Full text of "Records of the Auckland Institute and Museum"

See other formats

RECORDS
of the
AUCKLAND
INSTITUTE
& MUSEUM

Volume 29, 1992

a
a
.

yas
-

were | halite
aa af he eLTt

Ciel soem
ee > 'h
ats

+ gn
ah,
| ih sah

Rhy

“ag
on ¥

¥
oF he

f
le

="
ve

eee

Me te]
Cee ae
we

RECORDS
of the
AUCKLAND INSTITUTE
AND MUSEUM

Volume 29

AUCKLAND, NEW ZEALAND

Editors
B.J. GILL and N.J. PRICKETT

ISSN 0067-0464
RECORDS OF THE AUCKLAND INSTITUTE AND MUSEUM

Vol. 29 - 18 December 1992

Published by Order of the Council
G.S. PARK, Director

Auckland Institute and Museum
Private Bag 92018, Auckland 1, New Zealand

Printed by Prima Print Consultants Ltd, Auckland
Typesetting by Streamline Graphic Imaging Ltd, Auckland

RECORDS OF THE AUCKLAND INSTITUTE AND MUSEUM
VOLUME 29

CONTENTS

ARCHAEOLOGY

Additional information about the pa at Kauri Point, Birkenhead, Auckland.
Tate DaAveison- atid. RGp- WALLACE. SC... Levplaiberccdsame i pchoeweietameanweuie

Fibre fragments from the Raupa site, Hauraki Plains.

PPAR ERIN EL ARNE 8 oe is Scene SR TAIT e bua tes ted daen teed dae a eth eatin te cteet a peek

Archaeological excavations at Raupa: the 1988 season.

PSN MM 2 Fan) aoe eR IA ARATE ARAN i ears «PSB RESO OD Oe AONB PRMSERA: Pres NPR SOARES HPs 982 6.

PALAEONTOLOGY

A new lower Miocene species of Anodontia (Mollusca: Bivalvia).

WENT EEE i HE ae reels nade vies oe ee ion en cic micch seman aanraceninue euieteee

Paleontology and paleoecology of early Miocene sequences in Hays and
Tipakuri Streams, northern Hunua Ranges, Auckland.

MienAtr Facie-and ‘Bruce WAYWARD i Siciecccs pancdvtvnncadtadpeoauecdecssnenee

ZOOLOGY

Ecology and morphological variation in Placostylus bollonsi (Gastropoda:
Bulimulidae) at Three Kings Islands, New Zealand.

Fos cor amit a ule RuRTNTS ty ee OM eee ce, sh ec cane ae

The New Zealand species of Berothidae (Insecta: Neuroptera).

Weta. eee Pek re Se eR eR ie Bee ey DN en oS oe nae A ee

ADDITIONAL INFORMATION ABOUT THE PA
AT KAURI POINT, BIRKENHEAD, AUCKLAND

JANET DAVIDSON* AND Rop WALLACE**

* MUSEUM OF NEW ZEALAND
** ANTHROPOLOGY DEPARTMENT, UNIVERSITY OF AUCKLAND

Abstract. A radiocarbon date for the headland pa at Kauri Point, Birkenhead, is
reported. This suggests that the site was constructed and occupied relatively recently.
Analysis of charcoal recovered from the test excavations in 1971 indicates that the
site was in a cleared area surrounded by kauri-dominated podocarp forest.

The paat Kauri Point (N42/27; R1 1/35) occupies a prominent headland on the northern
shore of the Waitemata Harbour. It is an almost ideal setting for a fortified site, requiring
very little artificial defence, and is strategically located on the edge of the main harbour
channel overlooking access to the upper harbour and close to good fishing grounds. Despite
these advantages, the site appears to have been occupied only briefly. Test excavations in
1971, recently reported (Davidson 1990), revealed minimal structural evidence in the
interior of the pa and only a small amount of midden and other cultural debris.

THE AGE OF THE OCCUPATION

A sample of Chione stutchburyi shell was dated by the Nuclear Sciences Group, DSIR
Physical Sciences, Gracefield. The sample was taken from the fill ofa drain-like depression
at the back ofa slight artificial terrace in Area A (Davidson 1990: Figs. 3 and 4). The initial
determination was as follows.

NZA881 6 °C 0.64+0.1%
conventional radiocarbon age 840 + 120 B.P.

The calibrated age, as calculated by the laboratory, was A.D. 1290 to 1716 (95% confidence
level) and A.D. 1413 to 1627 (68% confidence level).

This sample was one of three submitted at the same time from different sites, all of
which were reported with unacceptably high standard deviations. Although the Kauri Point
date was older than expected, it was queried not for this reason, but because the size of the

Rec. Auckland Inst. Mus. 29: 1-5 1992

2 DAVIDSON & WALLACE

standard deviation rendered the result virtually useless as a means of placing the Kauri Point
site within the span of Auckland prehistory. The laboratory undertook to run all three
samples again. In the event, two further determinations were made on the Kauri Point
sample with the following results.

NZA1070 6 8C 0.41 +0.1%
conventional radiocarbon age 632 + 75 B.P.

NZA1735 6 °C 0.504+0.1%
conventional radiocarbon age 519 + 60 B.P.

The calibrated ages, as calculated by the laboratory, are as follows.

NZA1070 CAL. A.D. 1522 to 1894 (94%) 1938 to 1950 (1%)
1611 to 1815 (68%)

NZA1735 CAL. A.D. 1680 to 1950 (95%)
1696 to 1865 (68%)

The laboratory advises (Rodger Sparks, pers. comm. 1991) that the initial determination
(NZA881) should be disregarded, and the other two taken as acceptable indications of the
age of the sample.

These results suggest that the site was in use during the latter part of the Auckland
archaeological sequence. The calendrical age ranges of the dated sample do not exclude
the possibility of occupation well into the nineteenth century. However, the extent of the
stone assemblage (Prickett 1989) and the absence of objects of European manufacture older
than the late nineteenth century strongly suggest a late prehistoric occupation. It is likely
that the site was occupied during the eighteenth century.

Published traditional evidence about pa in the vicinity of Kauri Point is conflicting
(Davidson 1990:2-4). At least two different names have been suggested for the site in
question. It has often been thought to be Te Matarae a Mana, the pa of Manaoterangi, a
Kawerau chief who lived in the eighteenth century (Simmons 1980). The radiocarbon dates
do not resolve the confusion about the name of the site, but neither do they conflict with
the interpretation that this site was Te Matarae a Mana.

THE SURROUNDING VEGETATION

Nine bags of charcoal collected during the excavations were examined (by Wallace).
Most of the larger pieces were separated out from each bag and identified. Each piece was
snapped or cleaved along each of its three anatomical planes (transverse, tangential-
longitudinal and radial-longitudinal), mounted on plasticine on a glass slide, and examined
using a Carl Zeiss WL microscope fitted with an epi-illuminator based on a 150 watt Schott
fibre-optic light source. The charcoal was examined at 50 and 100 times magnification
under darkfield illumination and 200 and 500 times using plane polarised brightfield
illumination. Identifications were based on Meylan and Butterfield (1978) and Wallace's
own comparative material. The identifications are summarised in Table 1. Details of
individual samples are given in Appendix 1.

KAURI POINT 3

Table 1. Summary of charcoal identifications from nine samples of 87 individual pieces.

SSS
Species No. (%) of samples Total no. (%) of
species occurs in identified pieces
Manuka (Leptospermum scoparium) 6 (66.6) 45 (52)
Pohutukawa (Metrosideros excelsa) F (950) 25 (29)
Kauri (Agathis australis) 4 (44.4) 9 (10)
Matai (Prumnopitys taxifolia) 2 (22) 2 (2)
fern root (Pteridium esculentum) 1 (11) a
Towhai (Weinmannia silvicola) 1 (11) ae NY
Rimu (Dacrydium cupressinum) } (49 ] (1)

The fern root in this, as in other Auckland sites, is thought likely to be remains of growing
plants burned in situ, rather than accidentally charred fern root brought to the site as food.
The species present suggest three things: (a) the manuka and bracken an established
clearing; (b) the kauri, matai, rimu, and towhai a kauri-based climax forest to the landward
side; and (c) pohutukawa-covered slopes and cliffs on the harbour side.

Although the samples and identifications are too few for much to be discerned about
the distribution of species in different parts of the site, it is notable that whereas manuka,
and toa lesser extent pohutukawa are widely distributed, the podocarps are largely confined
to Area B. This was the area on the central ridge top from which most of the stone tools were
recovered. It was interpreted as a place where the inhabitants of the site gathered to work
at artefact manufacture and repair and possibly also to work fibre (Prickett 1989:192;
1990:18).

DISCUSSION

One of the most striking features of the limited excavations carried out on the pa at Kauri
Point was the very small amount of archaeological evidence revealed. The site appears to
have been only briefly occupied by people who left few traces of their presence. It has been
suggested that this is because it was built by people who came to the area for fishing, as oral
traditions relate, but that most of their activities took place down in the bay, the pa perhaps
serving only as a refuge and an expression of their mana (Davidson 1990;15). Soils in the
area are poor and it is unlikely to have been attractive for horticulture. Although a few pits
have now been discovered in the vicinity (K. Prickett, pers. comm.), these are the exception
rather than the rule in the field evidence of an area that presents a marked contrast to
volcanic zones of the Auckland region.

The charcoal identifications tend to support the interpretation that the general area was
still forested when the pa was built. It may have existed in quite a small clearing between
the harbour and the forest. In this case, it was probably occupied for brief periods during
the fishing season.

4 DAVIDSON & WALLACE

The difficulties experienced with the radiocarbon determination and the normal caveat
about reliance on a single sample, even without such difficulties, must make assessment
of the age of the site still tenuous. The two determinations NZA1070 and NZA1735
indicate a late prehistoric or early post-European age, while the nature of the assemblage
recovered from the site favours the former. Analysis of the stone assemblage suggested that
the site’s occupants lived under some stress and probably had difficulty procuring high
quality stone for tools (Prickett 1989:197). This would be quite likely during the troubled
times in the eighteenth century with which oral traditions relating to the area are concerned.

Acknowledgements, Janet Davidson would like to express her appreciation of the co-operation of the
Nuclear Sciences Group, DSIR Physical Sciences, Gracefield, in investigating the problems
associated with the radiocarbon date and the help and advice of Foss Leach in matters to do with
radiocarbon dating. Support for the dating and the charcoal identifications from the Science
Research Distribution Committee of the New Zealand Lottery Grants Board is gratefully
acknowledged.

REFERENCES

Davipson, J. M.
1990 Test excavations on the headland pa at Kauri Point, Birkenhead, Auckland, in 1971. Rec.
Auckland Inst. Mus. 27:1-18.

Mey ian, B. A., and B.G. BUTTERFIELD
1978 The structure of New Zealand woods. N.Z. Dept. Sci. Industr, Res. Bull. 222.

PrickeTT, K. E,
1989 The lithic assemblage from the headland pa at Kauri Point, Birkenhead, Auckland. /n:
Saying So Doesn’t Make It So, D.G.Sutton (Ed.). N.Z. Archaeol. Assn. Monog. 17:190-
198,
1990 Appendix 1. Obsidian analysis. Jn: Davidson (q.v.) 1990:16-18.

Simmons, D. R.
1980 | George Graham’s Maori Place Names of Auckland. Rec. Auckland Inst. Mus. 16:11-30.

JANET DAVIDSON, Museum of New Zealand, P.O. Box 467, Wellington. ROD
WALLACE, Anthropology Department, University of Auckland, Private Bag, Auckland.

KAURI POINT 5

APPENDIX 1. Charcoal identifications.

ee oun igenenities anmneiiealineenniiaie ene

No. identified pieces

Area A, square 1, fill of depression at base of bank
Pohutukawa (Metrosideros excelsa) l

Area A, square 2
Manuka (Leptospermum scoparium) 10

Area B, square 1, layer 2 (first sample)
Kauri (Agathis australis)
Pohutukawa
Bracken rhizome (Pteridium esculentum)
Towhai (Weinmannia silvicola)

NO WWD

Area B, square 1, layer 2 (second sample)
Manuka
Pohutukawa
Matai (Prumnopitys taxifolia)
Kauri

Ke py

Area B, square | extension
Rimu (Dacrydium cupressinum)
Matai
Kauri
Pohutukawa

~J Ree

Area B, square 3, layer 2, spit 2
Manuka
Area C, SW quadrant, layer |

Pohutukawa 10
Manuka . 2

Area C, SE quadrant (lower), layer 3
Manuka 14

Area D
Manuka 10

Kauri 1

FIBRE FRAGMENTS FROM THE RAUPA SITE,
HAURAKI PLAINS

MAUREEN LANDER

UNIVERSITY OF AUCKLAND

Abstract. Archaeological excavations at Raupa, near Paeroa, in 1988 uncovered the
charred remains of Maori fibrework dating from the early 19th century. Represented
are rain capes, plaited pieces, and possibly a korowai cloak. Interpretation has
depended upon similar fibre pieces recovered from dry cave environments in the
Waitakere Ranges near Auckland. The Raupa material adds to the small quantity of
Maori fibrework from securely dated archaeological contexts.

Excavations at the Raupa site (T 13/13) near Paeroa during January and February 1988
(see Prickett 1992), uncovered the carbonised remains of twined fibre, assumed to be
fragments of Maori rain-capes, and some small pieces of twilled plaitwork. These
fragments were subsequently conserved by Dilys Johns at the Anthropology Department,
University of Auckland, archaeological conservation laboratory. The fragments almost
certainly date from ca. 1820 and so offer knowledge of Maori fibre working techniques
during an important period of change some fifty years after effective European contact.

CONSERVATION OF THE FIBRE

The carbonised state in which the fibre was found had helped preserve the remnants
in the ground but made them very brittle so that they threatened to disintegrate once dry.
Some of the fragments were treated in the field with 6% w/w Rhoplex/water. Rhoplex is
an acrylic emulsion which was used to consolidate the surface of the damp fibre. These
pieces were then dried slowly in the field. In the laboratory, after cleaning, the fragments
were further consolidated with 7% Paraloid B72/acetone. Paraloid B72 is an acrylic resin
which penetrates into the structure of the dried fibre fragment to consolidate the whole piece
rather than just the surface. Material brought back as matrix and fibre for excavation in the
laboratory was treated with 6% Acrysol WS 24/water (dispersal emulsion) and then with
7% w/w Paraloid B72/acetone as above (Dilys Johns pers. comm.).

After treatment the fragments were reasonably firm and could be handled with care.

Analysis of the fragments was based on their appearance at this stage and from field notes
and diagrams provided by Dilys Johns.

Rec. Auckland Inst. Mus. 29: 7-23 1992

8 LANDER

ANALYSIS AND INTERPRETATION

Initial inspection of the fragments revealed differences in construction techniques and
treatment of the fibre indicating that the assorted pieces represented more than one garment.
Location notes and diagrams showed that a range of the material was excavated in
association with a large wooden bowl in a shallow pit in Area VII (see Prickett 1992: Fig.
17). These pieces included small remnants of twined kaupapa (groundwork of cloak), some
plain and some tagged on one surface, braided edges and collar fringes, short lengths of cord
and some small fragments of plaitwork. A more limited range of material was also found
in Area X (see Prickett 1992:Fig. 17).

Analysis of the Area VII material and the placement of fragments in relation to each
other led to the conclusion that they represent two garments, probably cloaks, which had
been inside and over the rim of the bowl, with another garment below the bowl and one or
more plaited items inside the bowl and below it. Here they were all burned in situ in a
shallow pit. One other garment was found, also burnt, in Area X. The fire accounts for the
fragmentary nature of the remnants and their carbonised state. The bow! (which may have
been damp or even have contained liquid) appears to have provided some protection from
complete disintegration.

Construction techniques

The term ‘twining’ is used in the context of this report to refer to the technique known
as whatu, traditionally used to manufacture Maori cloaks and some other items. ‘Plaitwork’
refers to the raranga technique commonly used to make kits, mats, belts and a number of
other articles (see Connor 1983:209, for a summary of differences between weaving,
plaiting and twining).

In order to identify and describe the material a selection of fragment samples were
divided into five groups to represent the five different items mentioned above. Selection
for each group was based on in situ location, construction technique and appearance of the
fibre. Fragments compatible with cloaks, for example plied cords, were included within the
groups with which they were found. Where the technique used was whatu (twining) the
material has been classed as ‘cloak fragments’ as a general descriptive term. This does not
exclude the possibility that they may represent more specific garment types such as maro,
piuptu, rapaki or pihepihe, all of which had features in common with rain capes and korowai
during the early 19th century (Pendergrast 1987:116-119).

Fragments in which the technique used was raranga (plaitwork) have been classed as
a separate group even though pieces were found in close association with cloak fragments.

Initial analysis of the material (Lander 1988) has since been modified and informed
by working closely with a range of fibre artefacts collected from caves and shelters in the
Waitakere Ranges and held at Auckland Museum (Lawrence 1989), Many of the features
present in the Raupa fragments were also evident in the Waitakere material. This helped
considerably in building up a coherent picture of items from which the Raupa fragments
may have come.

RAUPA FIBRE 9

Group 1: transitional rain cape/korowai?

The fragments in this group were excavated from inside and over the rim of the wooden
bowl in Area VII. They include braided neck edge pieces and a number of small pieces of
twined kaupapa (Figs 1-3). There is evidence of kokowai (red ochre) staining on some of
the material.

The whenu (warps) are of muka (flax fibre) which has been softened by beating with
a patu muka (stone pounder). This has given the fibre a slightly broken surface with less
distinct individual strands than unbeaten fibre.

The strands used as tag attachments on the kurupatu (collar fringe) have remnants of
leaf surface still adhering to them indicating that they were more likely to have been made
from strips of whole flax leaf which may have been lightly scraped. In time these may have
gradually broken down and become more fibrous, as seen in some raincapes in museum
collections. The Group | kaupapa pieces do not show any evidence of tag attachment to
either surface (Fig. 3).

The twining technique is aho patahi (single pair twining) using aho (wefts) of two-ply
rolled muka cord. The weft rows are spaced 15-20 mm apart and there are four or five wefts
per 10 mm. The whenu do not appear to have been twisted into two-ply cords. The whiri
(braid) on the neck edge looks like a five-stranded flat braid. There is a substantial amount
of red ochre on the inside of the neck edge and on the braid, but only small traces on the
outer fringe.

For cultural reasons, this image has been removed.
Please contact Auckland Museum for more information.

Fig. 1. Group 1: neck edge and kaupapa (base) - inside surface showing single pair
twining (whatu aho patahi).

10 LANDER

For cultural reasons, these images have been removed.
Please contact Auckland Museum for more information.

Fig. 2. Group 1: outside surface of the right-hand fragment in Fig.1. Flax tags have

been attached to form a collar fringe (kurupatu).

Fig. 3. Group 1: kaupapa fragment with no evidence of tag attachment.

RAUPA FIBRE | |

It is difficult to classify this group as remnants from a rain cape because of the absence
of tags on many of the fragments. From the evidence it appears to have been a kurupatu
and a plain kaupapa similar to one depicted by the mid-19th century artist George French
Angas (1972: Pl. 9). Another possibility is that it belonged to a transitional korowai type
in which tags similar to rain tags, or thrums of rolled cord, were attached intermittently over
the surface of the kaupapa for decorative effect. Transitional cloaks of this type were found
in caves and shelters in the Waitakere Ranges (Lawrence 1989:103-108).

From the distribution of red ochre on the inside of the fragments it appears that kokowai
may have rubbed off onto the cloak from the body of the wearer. Captain Cook visited the
Hauraki area at the end of November 1769 and left this description in his journal:

“The Natives resideing about this River do not appear to be very numerous, considering
the great extent of Country, at least not very many came off to the Ship at one time and
as we were but little a shore ourselves we could not so well judge of their numbers. They
are a strong, well made active people as any we have yet seen and all of them paint their
bodies with red Oker and Oyle from head to foot, a thing that we have not seen before.”

(Beaglehole 1955:210)
Group 2: rain cape

The fragments in this group were also excavated from inside the wooden bowl, but
nearer the bottom of it. They include pieces of tagged kaupapa and a short length of square
neck braid with fringe (Fig. 4).

The whenu appear to be unbeaten muka similar in appearance to the strands in the neck
braid which was probably formed from the ends of the whenu. After being taken into the
whiri the whenu ends were discarded to form part of the collar fringe. The flax tags on the
kaupapa are matted and fibrous with fragments of leaf surface still adhering to them (Fig.
5).

The twining technique is aho rua or double pair twining (Fig. 6), a method still favoured
by Maori cloak weavers. There are three whenu per 10 mm. The method of attachment of
the rain tags cannot be determined without damage to the fragments. A common method
of tag attachment is for the scraped muka ends of the tags to be incorporated into the kaupapa
as the work progresses. This method is often combined with adding doubled hieke-type tags
to the weft rows (Pendergrast 1987:85).

The square neck braid was of particular interest because I had not then seen anything
similar in the Auckland Museum collection. After much trial and error I was able to make
a reconstruction of the eight-stranded whiri and sent it to Diggeress Te Kanawa, a noted
Maori weaver, for her comments. She replied stating that neither she nor her mother,
Rangimarie Hetet, had ever seen or used this particular edge braid. She sent me back another
reconstruction which she had made in which the method differed from mine but the over-
all effect was similar. Later, when I was photographing the Waitakere material I came
across two more examples of the square braid, in particular on the cloaks from Karekare
(Fig. 7).

12 LANDER

For cultural reasons, these images have been removed.
Please contact Auckland Museum for more information.

Fig. 4. Group 2: neck edge with eight stranded square braid.

Fig. 5. Group 2: kaupapa piece showing tagged outside surface.

RAUPA FIBRE 13

For cultural reasons, these images have been removed.
Please contact Auckland Museum for more information.

Fig. 6. Group 2: double pair twining (whatu aho rua) on the inside surface.

Fig. 7. Detail of the eight stranded edge braid (whiri) on a flax tagged raincape found
at Karekare, Waitakere Ranges (Auckland Museum).

14 LANDER

Group 3: rain cape

The material excavated from below the wooden bowl was brought back to the
university as matrix and fibre for excavation in the conservation laboratory. The fragments
included some comparatively large pieces of kurupatu, smaller pieces of tagged kaupapa
and two short lengths of two-ply cord. There were also small fragments of raranga
(plaitwork) intermixed with some of the cloak kaupapa pieces (Fig. 8).

Initial investigations using a microscope showed that while most of the fibre appears
to be flax (Phormium tenax), there may be also some kiekie (Freycinetia banksii). More
stringent tests would need to be done in order to identify the fibre accurately.

Single pair twining (aho patah1) was used to construct the kaupapa which has the coarse
appearance and wide spacing between weft rows typical of many rain capes in museum
collections. The neck whiri is similar in neatness and tension to the other two found in this
area, perhaps all made by the same cloak maker. The kurupatu differs from the others
however in that there is no evidence of attachment to the kaupapa of the cloak. The method
of construction has not been determined as further investigation would damage the
fragments. It appears that lengthy doubled-over fringe strands were added into the whiri
during the braiding process (Fig. 9). The single weft row of aho patahi indicates that whenu
may have been incorporated into the whiri and discarded out again before being caught into
a final weft row and trimmed. If this was the case then the rest of the kaupapa has either
rotted off or been cut away. There was no evidence of it at the time of excavation (Fig. 10).
One of the Karekare cloaks (Lawrence 1989:108) has a kurupatu which had been stitched
on to the kaupapa (or perhaps repaired) with a running thread of plied muka (Fig. 11).

For cultural reasons, this image has been removed.
Please contact Auckland Museum for more information.

Fig. 8. Group 3: collar fragment of raincape intermixed with a small piece of plaitwork
(raranga).

RAUPA FIBRE 15

For cultural reasons, this image has been removed.
Please contact Auckland Museum for more information.

Fig. 10. Group 3: weft row of whatu aho patahi just below braid.

16 LANDER

It is probable that the tags were a mixture of single tags added in to the kaupapa, and
doubled over hieke-type tags added on top to increase the thatch effect. One fragment
actually shows evidence of the hieke-type tag attachment (Fig. 12).

The Group 3 fragments may represent a cloak similar to the hieke or tatara examples
in the Auckland Museum collection (Pendergrast 1987:85-86). The kurupatu remains a
mystery in that its method of construction and attachment have not been determined. It
could have been cut away deliberately and used as a separate piece over the top of the cloak.
I have discussed these possibilities with experienced Maori artists Digger Te Kanawa, Toi
Maihi and Paki Harrison, all of whom were as puzzled as I was by the pieces.

The tau (fastening cord) is made from two-ply muka (Fig. 13) rolled on the leg. The
Z-twist is formed by rolling the two strands separately in a move away from the body (if
using the right hand) and bringing the strands together in the movement back towards the
body. The tau may have been incorporated into the whiri for extra strength, as in the
Karekare example (Fig. 7).

For cultural reasons, this image has been removed.
Please contact Auckland Museum for more information.

Fig. 11. Kurupatu stitched or mended on to the kaupapa of a cloak from Karekare,
Waitakere Ranges (Auckland Museum).

RAUPA FIBRE 17

For cultural reasons, this image has been removed.
Please contact Auckland Museum for more information.

Fig. 13. Group 3: two-ply fastening cord (tau).

18 LANDER

For cultural reasons, this image has been removed.
Please contact Auckland Museum for more information.

Fig. 15. Group 4: edge braid with a small piece of knotted two-ply cord attached.

RAUPA FIBRE 19

Group 4: rain cape

A small selection of fragmentary fibrous material from Area X included tagged
kaupapa, a small length of whiri and a length of knotted cord, or tau (Figs 14-16). The
material is more fragmentary than that of the other groups and pieces that remain are matted
and decomposed. From what I can discern the material used in the kaupapa and whiri is flax,
prepared in the manner appropriate for rain cloak manufacture i.e. flax tags with scraped
muka ends incorporated into the kaupapa. The construction technique is aho patahi, similar
to the Group 3 fragments. There are three or four whenu per 10 mm and the weft rows are
spaced 30-40 mm apart. The whiri appears to be a five-ply braid, somewhat looser in tension
than the Group |, 2 and 3 braids, and so possibly made by a different cloak maker.

The presence of a small section of finer two-ply which has been knotted to itself (Fig.
15) may indicate that the kurupatu had been stitched or mended on to the kaupapa as
discussed earlier in reference to the Group 3 pieces. The tau (fastening cord) is four-ply,
made by rolling two two-ply strands together. The flax fibre used in the tau has a finer
texture and may have come from a different flax plant (some have finer fibre than others),
or have undergone some softening treatment - either pounding or rubbing.

Although the material evidence is somewhat scanty and decomposed there are enough
features consistent with rain cloak manufacture for the assumption to be made that this was
a rain cape of some type, perhaps similar to the Group 3 rain cape.

For cultural reasons, this image has been removed.
Please contact Auckland Museum for more information.

Fig. 16. Group 4: four-ply fastening cord (tau) which has been knotted in the centre.

20 LANDER

Group 5: raranga (plaitwork) fragments

An assortment of small pieces of raranga was found inside the wooden bow] in Area
VII (Figs 17 and 18). They represent one or more plaited items which had been with the
Group | and 2 cloaks at the time of the fire. A much smaller piece of raranga was intermixed
with Group 3 cloak fragments.

The material is flax leaf split to ca. 30-60 mm, Some may have been split again as work
progressed, as sometimes happens in raranga. The most intact of the pieces show twilled
plaitwork indicating a pattern of some kind in which strips pass over two or more strips of
the opposite orientation at a time.

It is not possible to say with any certainty what item or items the raranga pieces
represent. They could be remnants from kete (kits) or whariki (mats). However, there was
no evidence of edge braiding or handles to indicate kete, and the sparseness of the fragments
suggests items smaller than whariki. This does not rule out the possibility that they were
smaller pieces of matting which were sometimes cut up and put to a variety of uses, for
example wrapping.

From studying the items found together in the Waitakere sites it seems likely that the
Raupa raranga pieces could be the remnants of tatua (belts), although among Waitakere
material small pieces of matting were also found in association with cloaks. The Waitakere
examples (Lawrence 1989:110-113) were all made of twilled plaitwork similar to those in
the Raupa fragments (Figs 19 and 20).

CONCLUSION

Analysis of the Raupa fibre fragments has focussed primarily on interpreting the pieces
from a cloak-maker’s perspective. To this end I attempted to reconstruct one of the
techniques no longer used by cloak-makers, the eight-stranded neck edge whiri. This was
rewarding in that this knowledge was then able to be passed back to Digger Te Kanawa who
is a well known teacher of cloak making skills. For me this made the whole project
worthwhile: matauranga tuku iho na nga tupuna.

Initial research into the Raupa fragments was modified and reinterpreted after
comparison of the material with cloaks and other items found in the Waitakere Ranges. It
was then possible to build up a much better picture of the appearance of the Raupa items
when whole. The Raupa fibre fragments thus have provided a rich source of information
about the fibre working skills of the past.

Acknowledgements. I would like to thank the following people for their help - Dilys Johns, Caroline
Phillips, Pat Stodart, Nigel Prickett and Rod Wallace for information and assistance. Mick
Pendergrast for advice on weaving techniques, for showing me relevant cloaks in the Auckland
Museum collection, and for checking my findings. Digger Te Kanawa for her reconstruction of the
square braid and for sharing her skills and knowledge of Maori weaving.

RAUPA FIBRE 21

For cultural reasons, these images have been removed.
Please contact Auckland Museum for more information.

Figs 17-18. Group 5: plaitwork found in matrix with Group | raincape fragments.

22 LANDER

For cultural reasons, these images have been removed.
Please contact Auckland Museum for more information.

Figs 19-20. Plaited flax belt (tatua) found in a shelter at Piha, Waitakere Ranges
(Auckland Museum).

RAUPA FIBRE 23

REFERENCES

ANGAS, G.F.
1972 Portraits of the New Zealand Maori. Wellington, Reed.

BEAGLEHOLE, J.C.
1955 The Journals of Captain James Cook. The Voyage of the Endeavour. 1768-1771.
Cambridge University Press for the Hakluyt Society.

Connor, J.
1983 A descriptive classification of Maori fabrics: cordage, plaiting, windmill knotting,
twining, looping and netting. J. Polynes. Soc. 92:189-213.

LANDER, M.R.
1988 The Raupa Fibre Fragments. Unpublished research essay, Anthropology Dept., University
of Auckland.

LAWRENCE, J.
1989 The Archaeology of the Waitakere Ranges. Unpublished M.A. thesis, Anthropology
Dept, University of Auckland.

PENDERGRAST, M.
1987 Te Aho Tapu. Auckland, Reed Methuen.

PrIcKETT, N.

1992. Archaeological excavations at Raupa: the 1988 season. Rec. Auckland Inst. Mus. this
volume.

M. LANDER, Anthropology Department, University of Auckland, Private Bag, Auckland.

ARCHAEOLOGICAL EXCAVATIONS AT RAUPA:
THE 1988 SEASON

NIGEL PRICKETT

AUCKLAND INSTITUTE AND MUSEUM

Abstract. The second of two seasons work at Raupa, near Paeroa, Hauraki Plains, is
reported. A further 172 m? was excavated, to make up a total of 385.35 m? in two years
at the site. In 1988 work was concentrated on the area around a large house found in
1987, Work was also undertaken north of the house in what proved to be part
of the defences of the pa.

Two major occupation periods were identified. In the second (later) period an
intensive stone working area was located just east of the 1987 house. Artefacts
included a burnt wooden bowl and cloaks, large quantities of obsidian and chert, also
adzes, patu muka pieces, hoanga, nephrite chisels and ornaments, and tattooing
chisels and other bone items. Radiocarbon dates confirm archaeological indications of
18th and 19th century occupation only,

The lack of items of European origin and plentiful stone material in a major
settlement of 1820 indicate slow penetration of European material culture since first
contact half a century earlier.

The following report describes the results of archaeological investigations carried out
in early 1988 at the Hauraki Plains site of Raupa (T 13/13, formerly N53/37) on the left bank
of the Ohinemuri River at Paeroa (Fig. 1). This was the second of two excavation seasons
at the site. The 1987 work also has been described in the Records of the Auckland Institute
and Museum (Prickett 1990). The excavation took place for four weeks from 18 January
to 13 February 1988.

The Raupa work was undertaken in response to a proposal by the then Hauraki
Catchment Board to destroy the site in the course of flood control work along the Ohinemuri
River. Inthe event, the discovery of human burials at Raupa resulted in the Board foregoing
its ‘authority’ from the New Zealand Historic Places Trust and leaving the surviving part
of the site intact. Also left intact, and for the same reason, was the neighbouring site of
Waiwhau which was investigated by University of Auckland archaeologists in conjunction
with the work at Raupa (Phillips 1986, 1988; Phillips and Green 1991).

The report on the 1987 excavations includes an introductory section covering the
history, environment, archaeological history and research context of the site, which need
not be repeated here. It must be said, however, that there were two important contexts for
the work which were to some extent developed and modified for the second season’s work.

Rec. Auckland Inst. Mus. 29: 25-101 1992

O
as)
7
~

N
\

SOON

waihou RIVE!

SN \N

2 £5
n- SS WO WOT QW
« * SEN aK S
4 5 SS * 0
= : e

map.

Fig. 1. Location

RAUPA 27

The first was the description by Samuel Marsden of a visit to Raupa in the winter of 1820
(Elder 1930:255-256), and the second was the very large artefact collection from nearby
Oruarangi, a site similar in location and form, for which any work on nearby sites might
help give a more useful archaeological context. The Oruarangi collection also adds value
to any assemblage recovered from Raupa in describing the latter site.

In the first season’s work Areas I-V were excavated. Results included discovery of a
large house, a food storage pit, stone material dominated by a large quantity of obsidian and
chert, also adzes, patu muka pieces and a patu onewa blade, Other artefacts included a
pumice pot, fragments of fishhooks, bird spear, needle and tattoo chisels, and a hoanga with
incised design. Midden was not found in large quantities and was mostly crushed as a result
of being shifted about by the inhabitants of the Raupa settlement. A large number of post
holes, the distribution of stone material and other artefacts, abundant kokowai (ochre), the
location of midden deposits and some compacted occupation surfaces all helped in the
definition of activity areas and description of settlement arrangements.

Ambitions for the 1988 excavation were developed following the first season’s work,
The objective was to add to knowledge of the site, especially in terms of organisation and
use of space within the pa. Any additions to artefactual and economic information also
would be gratefully received. The major effort was concentrated in the vicinity of Areas
III and IV of the 1987 season, that is the area immediately about the large house found in
Area III (see Prickett 1990:78, 109). The intention was to further describe the use of space
about this important structure. In 1987 we did not excavate below the house so that it was
an important part of subsequent work to fully excavate adjacent areas and thus examine the
earlier occupation which we knew to be present from the 1987 Area IV and elsewhere on
the site. North of the house a 5 x 5m square (Area VIII) was excavated to examine an area
assumed to lie just behind the defences (Fig. 2).

The following account of the 1988 excavation is divided into three parts. Area VI is
described first, then the contiguous Areas VII, IX, X and XI immediately east of Area III,
and finally Area VIII to the north. The areas excavated were 47 m? (Area VI), 99 m? (VII,
IX, X and XI), and 26 m? at Area VIII, thus totalling 172 m?. There were also two machine
dug trenches opened up near Area VIII. In the two seasons work 385.35 m? were excavated
including 278.6 m? in the immediate vicinity of the Area III house. To put this into context
it is worth remembering that the site as a whole may have included a defended area of as
much as two hectares or almost 5 acres. Thus in two seasons work we examined ca, 2% of
the Raupa settlement area.

A brief account of the environment of the Raupa site is needed before the excavations
are described. The pa occupied a peninsula at the junction of the Waihou and Ohinemuri
Rivers. A narrow neck was cut by a series of three defensive ditch and bank earthworks
which could still be seen when the surveyor Courtenay Kenny visited the site in 1893 but
which were barely visible when Leslie Kelly was there halfa century later (Kelly 1945:207-
209). These defences were relocated by Caroline Phillips in 1984 by extensive machine
trenching of the northern part of the site which had been badly damaged by the Hauraki
Catchment Board taking fill for the renewal of nearby stopbanks a few years earlier.

28 PRICKETT

[| Area excavation
c 5 Trench excavation

www
7mm Stopbank

Fig. 2. The Raupa site showing Areas VI-XI excavated in 1988, also Areas I-V
excavated in 1987, and 1984 Trenches A and E. Inset shows relationship of
Raupa and Waiwhau sites.

RAUPA 29

Also taken for stopbank fill was water-laid silt, and rock flour from goldmining
operations upstream which involved the crushing of quartz bearing rock. Rock flour
overlay the entire site and filled the adjacent Waihou River channel, empty since the river
was re-channelled early this century. Deep deposits of silt in the site discovered during the
1987 excavations, above and below Maori occupation levels, showed that major flooding
episodes were not new.

Raupa occupied a natural levee at the river bank which was higher than the extensive
swamp of the Hauraki Plain lying all around. Marsden described it as “‘a point of high land”
(Elder 1930:255). An 1884 map (Anon.) showsa narrow strip of forest occupying the nearby
levee, with swamp behind. The swamp provided natural defence for a fortified settlement
which could be reached only by the river or perhaps by a narrow winding route along the
riverbank levees. In addition to the Raupa and Waiwhau settlements the levees were used
for gardening as noted by Marsden in 1820 (Elder 1932:256) and were otherwise clothed
in forest dominated by tall and straight kahikatea (Dacrycarpus dacrydioides) which so
impressed Cook some fifty years earlier (Beaglehole 1955:209). Raupa was a settlement
of the Ngati Tamatera tribe. The chief at the time of Marsden’s visit in 1820 was Te
Hikamate.

THE 1988 EXCAVATION
AREA VI

Area VI was opened up immediately south-west of 1987 Area III (see Fig. 2). The aim
was to give context to the house revealed in the first season and to excavate below the level
of the house in order to define any earlier occupation periods at this part of the site. At first
an area of 40 m? (10 x 4 m) was laid out; this was later increased to 48 m2 witha 4x 2m
extension at the north-west corner (Fig. 5). One of these additional metre squares (E1) had
already been partly excavated in 1987.

Stratigraphy

The upper part of Area VI was taken off by machine prior to hand excavation. Thus any
recent soil build-up and underlying rock flour and silt was largely removed. Beneath were
two major occupation layers described here as Layers 3 and 4. Area VI generally had 250-
600 mm of deposit. Detailed descriptions of the layers (for which see Figs 3 and 4) are as
follows.

Layer |. Blue-grey clay rubble from access road put across the Raupa site to assist Hauraki
Catchment Board work in the early 1980s (see Phillips 1986:98-99, 102).

Layer 2. Compacted yellow rock flour from goldmining quartz crushing operations in
the Karangahake Gorge, late 19th-early 20th century.

Layer 3. Sandy silt, mostly grey in colour but tending to brown above the dashed lines
on Figs 3 and 4 indicating a period of soil formation. Includes patches of sand
and compacted sand rubble, some stones, shell, charcoal and cultural material.

Layer 4. Brown clay, tending to black in places. Concentrations of stones in places.
Lies on disturbed surface grading into unmodified Layer 5,

30 PRICKETT

West
Datum A B C D tp

Layer 1 [Fon ie ee

: as Much red ochre
\ EGON

Fig. 3. Area VI north stratigraphical section. Layer descriptions in text. Note that the
datum level used in 1988 section drawings is the same as that in the report on the 1987
excavations.

White sand lenses

‘ Sy Se

Sand pra Shell LLL Fill from 1987 Area Ill

uM XA KK
Stones xxx xxx! Charcoal

Fig. 4. Area VI south stratigraphical section. Layer descriptions in text.

RAUPA 31

Layer 5. Yellow clay natural material.

The Area VI south section (Fig. 4) has Layers 2-4 as described for the north side of
the excavation. As at the north section, upper parts of Layer 3 tend to brown. At the base
of Layer 3 in the west half of the south section is a narrow charcoal band. Close to the east
end of Layer 4 are two discreet white sand lenses. A charcoal band lies at the base of Layer
4 in metre squares B-E.

The occupation sequence
In Area VI are two basic occupation levels relating to the Raupa settlement.

Level I. The first use of this part of the site (Layer 4) was for cooking, marked by a number
of shallow hangi scoops dug into the surface (Fig. 5). These scoops were dug into each other
with only the last of them retaining something of their original shape. The area was not as
intensively used for cooking as Area II 30 m to the south, which may have been
contemporary, or the north end of adjacent Area III which was used after destruction of the
house (Prickett 1990:103-104, 109).

Some hangi scoops were densely packed with fired cooking stones (see Fig. 6). Other
cooking stones were scattered about on the surface between the scoops, or, in two places
(see Fig. 5), had been brought together in small piles adjacent to hangi scoops where they
were ready for further use. Soil within and about the scoops was black and there was
charcoal and rare crushed shell (mostly pipi, Paphies australis) scattered about in places.

Level II. Spread over the Level I cooking evidence was Layer 3 consisting largely of sandy
silt with some material of cultural origin included in places (Fig. 7). This second cultural
surface had no evidence of cooking. A few postholes shown on the Fig. 5 plan probably
were dug from the Level II surface; most are in the eastern part of the excavated area.

The Level II surface relates directly to the Area IIT house floor excavated in 1987
(Prickett 1990:109). In Fig. 8 can be seen the west side of the 1987 excavation of Area III,
metre square Al2, with the previously excavated material visible from ca. 100 mm left of
the notice board to the metal peg and to a depth of ca. 150 mm. Thus, in Area VI, before
the house was erected in Area III, Level I cooking ovens were filled in and the ground
brought up to the same level as the adjacent house floor. The sandy silt of Layer 3 suggests
that flooding may have played a role in levelling up this part of the site. There is evidence
from the postholes for a light fence or shed here immediately behind the house but Area
VI Level II was largely given over to open space.

Radiocarbon dates
Analysis of two radiocarbon samples has been carried out by the University of Waikato

Radiocarbon Dating Laboratory. Both date the earlier Level I occupation. The location of
sample material is given in Fig. 5. Results are as follows:

32 PRICKETT

Cw 9

dooos i6buey

ajOyIsog O
wz L 0
N jc ES eS

‘ued | [9Ao] ‘TA Bory *¢ “SIy

L861 Ul |

peyeneoxg !
oO

(Z86L) Ill Belly

|
|
b jeung |

USPPIW ||9YUS

| asjdwes 19

RAUPA 33

For cultural reasons, these images have been removed.
Please contact Auckland Museum for more information.

Fig. 6. Area VI, Level I hangi scoop with oven stones.

Fig. 7. Area VI, surface of Level II.

34 PRICKETT

Sample 1: Pipi (Paphies australis)

Sample number: Wk 2039

8 °C: 0.8 0/00

Conventional radiocarbon age: 510 + 50 B.P.

Sample 2: Pipi

Sample number: Wk 2040

83°C: 0.7 0/00

Conventional radiocarbon age: 620 + 50 B.P.

The conventional radiocarbon age is based on the Libby half-life of 5568 years.

With application of the McFadgen and Manning (1990) correction for the marine
reservoir effect (Stuiver et al. 1986) these two radiocarbon ages give calendar years as
follows:

Wk-2039
Calibrated age range at 95% confidence: A.D. 1690-1950
Calibrated age range at 68% confidence: A.D. 1706-1867

Wk-2040
Calibrated age range at 95% confidence: A.D. 1533-1819
Calibrated age range at 68% confidence: A.D. 1574-1723

It is important to note that despite apparent differences in the two dates, statistically
they are of the same age. It is thus likely that the two radiocarbon dates refer to a single
occupation period. This reflects the comparatively thin Level I occupation surface and the
lack of cultural material, which together strongly suggest a brief occupation.

The two determinations may be pooled following Ward and Wilson (1978) to give a
mean radiocarbon age of 565 + 35 B.P. Corrected according to Stuiver et al. (1986) this
gives a calibrated age range for the pooled date as follows (6 R = -30 + 13, McFadgen and
Manning 1990):

Calibrated age range at 95% confidence: A.D. 1658 - 1829
Calibrated age range at 68% confidence: A.D. 1678 - 1728 57% probability
A.D. 1743 - 1792 43% probability

Adze

In metre square J5, Layer 3, was found the blade part of a once finely finished
greywacke adze (Fig. 23). The form is of a sharply defined shallow rectangular cross-
section, unlike any other found at Raupa but very similar to one from Oruarangi which is
illustrated by Fisher (1936: Fig.3, and see measurements given on p. 18) and included
among his Type B. The fragment is highly polished overall except for a small line of
hammer-dressing down one margin. The severely battered blade is 54 mm in length and
the maximum depth of the recovered fragment is 15 mm.

RAUPA 35

For cultural reasons, this image has been removed.
Please contact Auckland Museum for more information.

Fig. 8. Area VI, east section of metre square E1 showing location of Burial |

excavated in 1987 at the north (left) end. Approximately 100 mm left of the notice board

a vertical line can be seen to a depth of ca. 150 mm, this marks the Area III house floor.

The pale fill beneath lies on top of the Level | surface which is exposed in the
foreground. Scale has 250 mm intervals.

Nephrite

Also from metre square J5, Layer 3, came the broken fragment ofa nephrite chisel (Fig.
34). The chisel was highly polished to a rectangular cross-section 11 x 7 mm, the remaining

butt end fragment being 46 mm in length. The stone is pale green pounamu or nephrite.
Patu muka

Two patu muka handles were found in Area VI: both were assigned to Layer 3 and thus
relate to the Level II occupation period. Together with two items found in Area IV in 1987
and other fragments from Areas VII and X reported below, a total of six patu muka are
represented by pieces recovered during the Raupa excavations.

From metre square C5 came the well formed handle and part of the main body of a
massive patu muka (Fig. 24). The raw material is a fine-grained andesite such as makes up
a large part of the nearby Coromandel Range and peninsula. This item was not broken
through use but is clearly fire cracked, either accidentally or asa result of the already broken
patu muka being used as an oven stone and suffering more damage as a result. On one side

36 PRICKETT

is evidence of kokowai grinding with some hammer-dressing over part of the red surface
to indicate subsequent use as an anvil, perhaps for the breaking up of kokowai lumps.

The second item, from metre square E5 is a handle end, fashioned from a medium-
grained andesite (Fig. 25). Unlike the other handle, the butt end found in Area X (see Fig.
26) and the handle found in Area IV in the first season (Prickett 1990:146), in this case there
is only the slightest knob at the handle end. This minimal shaping of a patu muka from a
water-rolled stone of the right basic shape is common also among Oruarangi examples now
in the Auckland Museum.

Obsidian

A total of 151 pieces of obsidian weighing 862.1 g was recovered from Area VI (Table
1), The average weight per square metre is 18.34 g for the fully excavated 47m. More than
80% of the total was assigned during excavation to Layers 1-3, almost all of it to Layer 3,
i.e. to the upper occupation level. Lower occupation level (Layer 4) material was largely
confined to the eastern part of the excavated area.

Approximately 70% of pieces were less than 5 g in weight, of which almost half were
tiny fragments less than | g. Of the total number of flakes, 14.6% weighed 5-10 g, 8.6%
10-12 g and 6.6% more than 20 g. Outstanding was a 79.9 g flake of grey obsidian, probably
from the Waihi source, with some water-worn cortex to indicate an alluvial ori gin. Another
large flake, this time of green Mayor Island material, weighed 76.6 g. It displayed possible
edge damage in the form of a notch on one side.

More than half the obsidian pieces (ca. 53%) showed no sign of use from cutting or
scraping tasks. One-third of these were classed as shatter waste, and the remainder waste
flakes struck accidentally or on purpose from a larger block. Of the 47% displaying
evidence of use in one form or another, ca. two-thirds had uni-facial edge damage or
possible edge damage. Some of these items may have been broken from a larger piece after
use as a tool; for example, a 0.2 g flake with edge damage from metre square J5 can hardly
have been used on its own. Most of the remainder displayed a mixture of battered, edge-
damaged and notched margins. A few pieces have the appearance of core remnants from
which flakes were struck when needed.

Almost all the obsidian is green in transmitted light and in all probability originates
from Mayor Island. Among this material are some pieces which have original cortex from
water-rolled boulders or flow margins. Of other material, four pieces totalling 82.6 g have
been identified as originating from the Waihi source (Moore and Coster 1989), with another
seven of pale grey material (total 7.0 g) indicating a central North Island or Coromandel
source. Raw material percentages by weight are: Mayor Island 89.6%, Waihi 9.6% and pale
grey 0.8%.

There is little to be said of the distribution of obsidian in Area VI. Layer 4 material
is strongly concentrated on the eastern half of the excavated area and may reflect only one
or two episodes of use. More abundant Layer 3 obsidian is widely scattered, with metre
Squares of greater than average weight occurring commonly to the south and west.

RAUPA 37

Table 1. Summary of Area VI obsidian distribution; ofa total of 151 pieces of 862.1 g. Refer
to Fig. 5 for metre square locations.

Square No. Layers 1-3 No. Layer 4
Wt (g) Wt (g)

Bl 1 0.9

Cl Bs rae:

D1 2 0.7 1 10.4

El(see Area III square A12, Prickett 1990:119)

B2 1 3.9

C2 4 39.2

D2 2 3.2

E2 4 10.6

A3 7 eT.7

B3 1 0.2

E3 1 4.6

F3 4 52.8 2 4.4

G3 z 4.6 1 ie

H3 2 18.2

13 : 10.5

J3 4 18.1 2 9.5

B4 9 9.4

D4 1 28.3

E4 1 5.1

F4 3 233

G4 5 53.4

H4 1 2.6

14 10 53,9 5 y AN

J4 9 31.0 l Pay |

AS 9 47.9

BS 5 43.2

cS 1 17.4

E5 3 2.0 ] 76.6

FS 3 8.5

G5 0.7

H5 8 224 l 4.5

IS 5 2.6

J5 l 0.3 0.2

A6 & 8.6

B6 2 3.6

C6 30.2

E6 4 122.8

F6 1 0.6 e 11.6

16 1] 0.3 2 Si:

J6 5 7.8 1 3.0

129 694.0 22 168.1

38 PRICKETT

Chert

Ninety-eight pieces of chert weighing a total of 2054.7 g were recovered from Area
VI. Nine pieces (120.7 g, 5.9% of the total) were assigned to Layer 4 (Level I) during
excavation; the remainder and vast bulk of material came from Level II.

Table 2 outlines the distribution of chert by metre square and layer (see Fig. 5). The
distribution is in very marked contrast to that of obsidian. The latter is most strongly
represented in the eastern part of the excavated area while chert is concentrated at the
western end. From only five metre squares (C4, E4, B5-6, D6) was recovered 1325.6 g or
71% of the total Layer 3 chert. Ifeight adjacent squares are added then these 13 m?, a quarter
of the whole, account for 88% of all Level II chert.

Cherts from Area VI include a wide range of colours and quality. Most is the same
relatively coarse-grained white or creamy white stone that was predominant in Areas I-V
excavated in 1987 (see Prickett 1990:91, 120, 133). This stone can vary to pink, red or
earthy yellow colours and may be of chalcedonic quality in places. A 326.7 g block of
uniform cream chert of this type was found in metre square C4, Layer 3. Other large pieces
weighed 206.7 g (E4) and 286.0 g (B5). Altogether, in excess of 80% of all Area VI chert
is of this raw material, which was almost certainly available from chert bearing deposits
in the nearby Coromandel Range.

Other cherts were varied. Two fragments of jasper with white hydrothermal veining
and black water-rolled cortex which totalled 35.7 g, were found in square B6, Layer 3. A
piece of cream chert with grey and ochre coloured veins from square D6 weighed 113.5 g.
A number of small pieces of high quality fine-grained white or creamy white chert were
found, including two fitting pieces (27.3 g) from F4 which display some previous and
subsequent flake removal and also possible edge damage through use. A 12 g piece of good
quality creamy white chalcedonic chert probably relates to the predominant typically
coarse chert which occasionally includes chalcedonic quality material. (An example of the
latter is given by an 81.6 g flake of creamy chert from metre square D5 which grades over
a few millimetres from a material almost of sandstone coarseness to a translucent
chalcedony.) Some pieces have fine striations on one or more faces which are similar in
appearance to petrified wood. Other material represented includes poor quality opaque pale
grey and fine-grained black and yellow cherts.

Few pieces of chert show any sign of use. More than 75% are shatter rather than
deliberate or even accidental flakes. As many as 15 pieces have evidence of deliberate
flaking either before the recovered flake was struck or after. Only nine pieces display
possible edge-damage as a result of use, with most of this evidence very slight, perhaps even
having occurred during or after excavation. Three large pieces may have provided a
resource block of material but there is no sign of use as a core. The contrast is marked
between the high rate of utilisation of obsidian and the very low rate in the case of the chert.

Kokowai

Kokowai or red ochre was present in large quantities throughout Area VI, in both
Levels I and II. Much of it was in the form of tiny pieces which left a smear when cut by

RAUPA 39

Table 2. Summary of Area VI chert distribution; of a total of 98 pieces of 2054.7 g. Refer
to Fig. 5 for metre square locations.

Square No. Layers 1-3 No. Layer 4
Wt (g) Wt (g)

Bl 2 1.4

D1 1 0.4

EI ] 2205

E2 3 119

A3 1 2M

D3 2 a4 5 DEES

J3 4 oar

C4 2 SO

D4 4 37.4

E4 2 209.3

F4 2 Be ae:

14 3 5.0 Z 1

J4 4 20.3 1 1.8

AS 9 ss

BS 4 356.9

Go ] A

DS 1 81.6

E5 ] 28.5

F5 2 22.4

G5 2 10.2

H5 3 14.5

IS 4 7.6

J5 2 4.0

A6 2 75.8

B6 8 159.5

C6 4 34.0

D6 6 226.2

E6 6 38.9

F6 3 14.8 l 3.9

89 1934.0 9 120.7

AQ PRICKETT

the trowel during excavation. The area was also notable, however, for the comparatively
large number of substantial pieces of kokowai, some of it of high quality.

Red and yellow ochre were both present, much the greater part being red, ranging from
orange/red to a dark wine red/brown colour. The yellow ochre was ofa pale but pure yellow
colour, much of it, like the red material, still in its unmodified rock form of soft gritty
material.

Most of the 25 ochre pieces large enough to be collected came from the southern part
of the excavation (metre squares 5 and 6). All except one were assigned to Layer 3 during
excavation. The remaining piece came from Layer 4. Also from the southern part of the
excavation came six hoanga (grindstone) pieces, some of which had been used for the
grinding of kokowai, as indeed the patu muka handle mentioned above had been used as
a platform in ochre grinding.

Hoanga

Six hoanga or possible hoanga fragments were found in the southern part of Area VI.
All of them are associated with Layer 2 or 3 and thus probably belong to the second
occupation period which relates to the Area III house. It is possible they originated in the
earlier occupation and were brought in with fill to level up this part of the site; however,
the fact that kokowai pieces also largely come from Level II and also from the southern part
of the excavation area suggests that hoanga and kokowai originated here and belong
together.

The most important item is made up of two fitting pieces of a flat andesite slab hoanga
measuring 315 x 205 x 30-40 mm (Fig. 48). The pieces were found in adjacent metre squares
A6 and B6. The flat under surface of the B6 piece has been altered by fire since the hoanga
broke in half. Before this, the upper surface of the intact hoanga was dished out as a result
of grinding to expose the coarse hard rock beneath. This oxidised material is of the right
colour and consistency to make up a good quality kokowai and it may have been used for
this. Alternatively the fine soft patina material may have been used as a grinding medium
against the hard surface beneath. This item has some small areas of kokowai ground into
exposed rock but it does not appear to have been used extensively for kokowai grinding.
Two pieces of very high quality fine-grained kokowai were found in close association with
one half of it - a brick red piece of ca. 80 g and a deeper wine red pebble of 7.6 g.

Three small fragments from metre square E6 clearly have been broken off a single
hoanga. They measure 54 x 24 mm, 47 x 26 mm and 26 x 20 mm respectively. The raw
material is a natural andesite slab. Each piece is marked on the intact flat working surface
by finely ground brick red kokowai.

A last item also comes from square E6 and again derives from a natural slab of coarse-
grained andesite. It is unclear if it has been used as a grindstone. It has been broken from
a larger piece and now measures ca, 120 x 95 x 30 mm, The flat upper surface is covered
in soft brown patina and some areas of black charcoal.

RAUPA 41

Other stone

Apart from the large quantity of oven stones there was very little stone material other
than that introduced above. A 930 g fragment of greywacke (metre square G4) was probably
shattered by fire off the large water-rolled boulder from which it came. Such boulders were
much too large for oven stones and most likely were imported as anvils for the beating of
fernroot (see for example Fig. 18).

A piece of petrified wood (square F3) 83 x 31 x 18 mm has some red chalcedony at
one end. Such material was also found in the 1987 excavation, notably in Areas III and IV
(Prickett 1990:123, 134). The raw material was sometimes used for lure shanks where
longitudinal strength was needed. Two small pumice pebbles also were among recovered
material; both are unused and probably arrived naturally at Raupa.

Kauri gum

Some small fragments of highly degraded kauri gum were found in metre square B4,
Layer 3. It was also recovered from Area I in the 1987 excavations (Prickett 1990:94), and
the comment may be repeated that kauri gum was used in one method of preparing tattoo
pigment.

Material of European origin

A hollow iron tube 3 mm in diameter with encircling (but not screw threaded) grooves
on the outside may be 10-15 mm in length. The lump of sand and other material which now
encases it as a result of iron oxidisation renders most of the item invisible, Although it was
assigned to Layer 3 when excavated it is most likely this object post-dates the Raupa
settlement.

AREAS VII, [X, X and XI

At the east side of the 1987 Area III, south of Area 1 V, excavation was initiated in 1988
with the opening up of a 30 m? (10 x 3 m) Area VII, 15 m? (5 x 3 m) Area XI and 18 m (6
x 3 m) Area X (see Figs 2, 9). Later Area X was extended to 30 m? (10 x 3 m) and Area X1
of 24 m? (6 x 4 m) was opened up between VII and X. Thus the area excavated in this part
of the site in 1988 totalled 99 m?. To this an be added Areas III (107.6 m2), IV (25 m?) and
VI (47 m*) to make a total contiguous excavated area here over the two seasons work, of
278.6 m?. Areas VII, [X, X and XI are dealt with together for convenience.

Stratigraphy

Stratigraphy over the combined Areas VII, IX, X and XI was relatively complex, with
different materials making up the main occupation levels in different parts of the
excavation. In particular Layer 3, the upper occupation layer, varied considerably. Over
most of the northern part of the excavation it was represented by a homogeneous grey sandy
silt, similar to that encountered in Areas III and IV excavated in 1987 (Fig. 12). Towards
the south, however, it was represented by a mass of shell midden and black soil (Fig. 13).

42 PRICKETT

For cultural reasons, this image has been removed.
Please contact Auckland Museum for more information.

Fig. 9. View of excavations: Area VI front right; Areas VII, IX and X under excavation
beyond with Area XI not yet opened up between them. Trees beyond mark the bank of
the Ohinemuri River.
South 10 9 DATUM 7 6

me ape aS ‘ A EWS ‘ as SOS SS NS Sd RS Sy eG
\ ‘ \ ON Re De Boe es SOU. ING

a SES . Sia %, ne Rea EWN LEN Xs MQQQXL ~ \ a

5 4 3 2 1 North

os ‘ of s, 2 \ \ x rs \ \ \ \ We rhe ~ Se, = c x, \ a Rye “. \ \ ~e \
Roce Ba ty ae i ies NA Ve SNS So ee PERS 2 ae \ ‘ Set et RDS Su AS \
ras Senter te BRE OR. NS ONE Se gy NS SAH

Oo 1m mR 7
dx Charcoal

Fig. 10. West section of Areas X, XI and VII (south to north). Layer descriptions in text.

RAUPA 43

There were also ash and sand lenses. In the following description (and in Figs 10 and 11)
different material, which nonetheless belongs to the same stratigraphical unit, is given the
same general layer number.

Layer 1. Surface fine yellow rock flour mine tailings, left after machine stripping of
surface prior to excavation.

Layer 2. Mottled crumbly brown loan. This material was found throughout the Raupa
site. In adjacent Area IV it was also labelled Layer 2 (see Prickett 1990:124-
125), and in Area ITI, Layer 3 (Prickett 1990:110). Over some of Areas VII, [X-
XI it was missing, probably having been stripped off during removal of fill for
the early 1980s stopbank work,

Layer 2a. Brown-grey silty loam pit fill.

Layer 2b. Yellow clay lens in pit fill.

Layer 3. The Layer 3 complex relates to the later (Level I]) occupation and includes a
large number of layers and lenses. All are designated Layer 3 to simplify
reference to Level II material. Layer 3 relates to Layers 3 and 4 in Area IV and
to the Layer 4 house occupation surface in Area III (Prickett 1990:110, 125-126,
110). The main body of material (“Layer 3”) is a homogeneous grey sandy silt
which includes charcoal and shell fragments.

Layer 3a, Compact dark grey sandy clay loam. This layer is identical to the house floor
and related occupation surface of Areas III and IV (Fig. 10).

Layer 3b. A deposited mass of midden, mostly shell, is largely confined to Area IX and
the adjacent eastern end of Area X. In Squares L-M and the lower part of Squares
I-K the midden is comprised of dense shell, above this in I-K is some less dense
midden in a sandy matrix (see Figs 10 and 12). Charcoal is abundant towards
the eastern end. Within Layer 3b are also two distinct sandy lenses and an ash
lens (Squares I-K).

Layer 3c. Grey sandy silt with some clay mixed in and fragments of charcoal and shell.
This layer is made up of Layer 3 with a slight admixture of underlaying Layer
4, the natural base of the site.

Layer 3d. Yellow and grey mottled silt/loam pit fill.

Layer 3e. Pit fill made up of banded brown sandy silt loam,

Layer 4. Yellow clay natural base of the site.

The occupation sequence

Stratigraphical evidence indicates two major periods of occupation in this part of the
Raupa site, as in adjacent Area IV (Prickett 1990:126-128). Over much of the excavation
it was difficult to sort out Level I and II features, because of the uniformity of the material
which was being excavated. Many postholes were recorded on field plans of both levels;
they are excluded here from the lower Level I plan as clearly having originated from the
Level II occupation surface where they were first noted. A number of graves and pits in the
area date from before and after final occupation of the site.

Level I, First occupation of this part of the Raupa site included areas of midden with
charcoal and shell, ash lenses and many post and stakeholes (Fig. 14). Through it had been
dug several pits which contained burials or fragmentary human bone.

44 pRICKETT

East M L DATUM J 1

B A

Mia Ast ae Ss eee Bae

LL [ELL ee Lita 7
{iY N

BN Sand
SO SN \

Fig. 11, South section of Areas IX and X (east to west). Layer descriptions in text.

Postholes and hence wooden structures, are concentrated across the southern part of
the excavated area. The most interesting group is in Area [X (squares K, L and M) where
there is evidence of a rectangular structure, in all probability a dwelling house, ca. 2.5 m
wide and 4 m long or more. From the archaeological evidence and from historical
observations that Maori houses generally faced between east and north (Prickett 1982:129,
131) we probably have the rear and side walls. Surviving timber fragments in what would
then have been the rear ridge posthole have been identified as totara (Podocarpus totara).

Within the building was found a most unusual feature which may be described as a
clay-lined bowl or pit (Fig. 15). This gave rise to some speculation as to its purpose so that
it deserves detailed description here. The clay bowl was set 280-320 mm into the ground
penetrating 60 mm into the natural base of the site. It was oval in plan measuring 550 x 480
mm, its long axis extending north-north-east/south-south-west. The well formed rim was
40-60 mm wide while lower down the sides and at the base of the bowl the clay was 90 mm
thick, to as much as 130 mm at the angle between the near vertical wall and flat base. At
some stage in its history the bowl had been relined: the original grey clay with rust brown
inclusions and rare pipi fragments was covered by a veneer or crust of brown/black colour
over which had been plastered 20-50 mm more of grey clay.

RAUPA 45

For cultural reasons, these images have been removed.
Please contact Auckland Museum for more information.

Fig. 12. Area VII south section. The Level I occupation surface can be seen in the

foreground. The Level II surface is clearly visible in the section right of the noticeboard,
but above and left of the notice Layers 2 and 3 are mixed without a clear Level

section showing Lay as Shallow lenses of shell midden cut
by a Level II ditch below the notice board.

46 PRICKETT

N \ @ |
om \ e,° oO e
. bite hy ie . Og
kahikatea ¢ ee ead ° —
\ / #

Ne
A
> ° ; & oO O O <
a © aes, O 4
re ond a _ O l
\ a |

@ \? X

etn |
O 545 |
10 e |
+ ol
|

Fig. 14. Areas VII, IX, X and XI Level I plan.

RAUPA 47

Posthole >250mm

Posthole

{

A
O @\%

Soe

Depression

Charcoal

Shells

ae
( He

Stones

48 PRICKETT

For cultural reasons, this image has been removed.
Please contact Auckland Museum for more information.

Fig. 15. Area IX, metre square L8. Clay-lined pit with fill half-sectioned and clay cut
down toa depth of ca. 100 mm. The shell from overlying Layer 3 can be seen in the upper
part of the fill.

There can be little doubt that this unusual feature relates to the house which encloses
it. Itis ca. 1.5m from the rear wall and midway between the two side walls. The orientation
of the long axis of the pit conforms exactly to the house orientation. It is thus located as a
fireplace would be - on the house centreline and near the rear wall. Itis, nonetheless, a most
unusual hearth. Nor is there direct evidence that it was used as such; it contains no ash or
charcoal, being filled with 60-80 mm of water-laid dark grey sand, over which was sand
and charcoal and, above that, shell from overlying midden material. Within the fill was
found a single human wrist bone (right triquetral). Nonetheless, that this feature relates to
the house and that it was a hearth seems to be the most economical interpretation.

West of the Area IX house, and confined largely to Area X are postholes which mark
several structures. Many of the postholes are paired to indicate replacement. One line of
deep holes (greater than 250 mm, see Fig. 14) extends east-west to the corner of metre square
F7. From the central pair of deep postholes in this line another line with several paired holes
can be traced extending to the south-east. It seems likely these represent a roofed structure
as much as 6 m in length and 3 m wide, the remaining evidence for which lies in the
unexcavated area to the south. Other groups of postholes which are even less open to
interpretation lie between the intrusive Burial Pits 3 and 8 in Area XI, and at the east end
of Area VII.

RAUPA 49

Shallow depressions were found in several parts of the site, notably a group filled with
fine ash between the Area XI burial pits and another group in squares G and H, Area VII.
Patches of charcoal and some midden in squares M6-7 (but which may properly belong with
overlying Level II material) completes the Level I evidence.

Level IT. The upper occupation level is contemporary with the Area III house excavated in
1987 and Level II in Area IV (see Prickett 1990:109, 127). In some parts of the excavated
area the late occupation level includes more than one event or activity (Figs 16 and 17).

Only at the west part of the area was there evidence of a compacted surface such as
was present over the Area III house floor and the adjacent part of Area I'V (see Fig. 10). Much
of the upper occupation surface was of relatively soft silt loam on which it was not always
easy to define the occupation surface (see for example Fig. 12).

Postholes and their surface structures have a very different distribution to those of
Level I. Notable is a building in Area VII which included a number of posts identified as
totara and kauri (Agathis australis) from in situ fragments. Some of the pieces of wood were
of light dressed planks similar to evidence found in adjacent Area IV in 1987 (Prickett
1990:126). There is a problem here in that in Area IV the dressed planks were assigned to
Level I. Itis likely, however, that they do belong to the same structure possibly a dwelling
to judge from the attention paid to the well-finished posts. The distinct line of posts at the
east side of Area IV (Prickett 1990:125) can be seen to extend into Area VII, square El,
before a right-angled corner to the line which includes the kauri and totara posts extending
to the east. The difficulty of assigning posts to an upper occupation layer is illustrated by
Fig. 25 in the 1987 report (Prickett 1990:114) which shows a trench dug as much as 100
mm below the house floor in order to find some of the associated posts.

At the west side of the excavated area are a number of postholes possibly belonging
to two or three lines extending south-west/north-east across Areas X and XI. If there was
a roofed structure here as the evidence suggests it must have been very close to the Area
IIT house. In fact it would have been close to the position of a building immediately east
of the larger house, as is shown in the artist’s reconstruction of the site prepared from
evidence of the 1987 work (see Prickett 1990:149).

In various locations about the excavated area were small groups of stones, some of
them with charcoal, andin Area IX one with shell also. Such grouped stones are reminiscent
of those in Area VI and other cooking areas where they are associated with hangi scoops,
which were, however, absent here.

Much larger boulders found in Area VII had an altogether different purpose. These
well rounded river boulders weighing 8-12 kg had been brought to the site, probably from
the Karangahake Gorge of the Ohinemuri River, a few kilometres upstream. They were
found in three locations in Area VII (see Fig. 16) where there were one, two and three stones
respectively. One stone was carefully wedged in place by a spawl and smaller stones (Fig.
18). These boulders served as anvils probably for the pounding of fernroot, or possibly also
for the beating of flax to soften the fibres for use. This part of the Raupa settlement at the
side of the Area III house was clearly an important area for such activities.

50 PRICKETT

Fig. 16. Areas VII, IX, X and XI Level II plan.

RAUPA 51

(

N
A c
42.
YRS, %

fo)

e 2 ©

- +
oOo on Q
Oo oO ®
Oo 1a (a)
ea
O®@8 .»
ele

x Xx

x x

x
xx Charcoal

Oye

=
fe)
a
>
=
re)

52 PRICKETT

For cultural reasons, this image has been removed.
Please contact Auckland Museum for more information.

Fig. 17. Area VII Level II surface from west. Excavation of burnt bowl and adjacent
burial pit 7 has just begun.

RAUPA 53

Two important finds, near the west end of Area VII and in Area X, were of fragmentary
plaited and woven textiles, The material is described in detail by Maureen Lander, also in
this volume, and needs only to have its archaeological associations described here.

The Area VII, squares B2-3, find was in a shallow scoop into which had been thrown
a wooden bowl made of totara (Fig. 19). Mostly inside the bowl, but also outside and beneath
were the burnt fragments of three woven garments and some plaited flax fibre which may
have come from a mat (whariki). All had clearly been burned in situ as the extreme fragility
of the fibre would not have allowed any movement after the fire. The burning had allowed
the preservation of valuable information since the wood and fibre would not have survived
but for being totally charred.

In Area X there was a smaller find of charred fragments, probably of one garment.
While the Area VII material suggested preparation of a fire in situ, it is possible the Area
X find was a burning fragment which originated elsewhere and was thrown aside before
burning was complete. The Area VII fire marks the deliberate destruction of several items
which might be considered valuable in terms of the time taken to make them and possibly
also in terms of the prestige assigned to them. Ash scattered from squares G and H2 down
the east side of Area XI and into Area X may be associated with the shallow fire scoop visible
in Fig. 12.

Throughout Area IX and at the east end of X was an extensive area of dumped midden.
Many pipi and cockle shells still had the two halves joined to show that this material
probably lay where it was initially dumped after use. The midden is almost certainly part
of the same extensive deposit which appears as a dense layer in Trench E (see Fig. 2) 10 m
to the south.

Also scattered over the surface of Layer 3 were fragments of human and dog bone. As
in Area III it was scattered over and in some cases trodden into the occupation surface.
Because this was mostly very fragmentary and because some was clearly dog it was brought
back to Auckland to undergo detailed identification and examination, for which I am
indebted to Elizabeth Hudson.

The mid-section of an adult human palate with some tooth sockets (but no teeth), was
found beneath a 170 x 100 mm broken water worn stone in metre square E5. From the same
area came fragments of burnt human adult cranial bone scattered in D4, E4 and FS. Clearly
it had been burnt in an open fire before being scattered over this part of the occupation
surface. Fragmentary human teeth from squares E5 and A7 may be associated although
there was no indication of burning. In metre square B2 was found the top part of a human
skull with dog mandible and maxilla fragments nested inside. From B5 came some adult
foot bones. In addition to the dog bone found with the human skull piece in square B2, most
of a dog cranium was recovered from square D5.

Burials
There were three human burials and four pits with some human remains in Areas VII,

X and XI. These displayed a variety of burial practices and need to be looked at in turn
before general comments can be made. Their locations are given in Fig. 16. Burials | and

54 PRICKETT

For cultural reasons, these images have been removed.
Please contact Auckland Museum for more information.

Fig. 18. Area VII, metre square C2, anvil stone. A spawl is wedged at the side to hold
it in place.

Fig. 19. Area VII, metre squares B2-3, showing the rim of the burnt wooden bowl. Scale
intervals at 250 mm.

RAUPA 55

2 were excavated in 1987 and are discussed in that season’s report (Prickett 1990:117).
Identification of pits 6, 7, 8 and 9 as burials depended on occasional or fragmentary human
bones and plentiful ochre, along with their general resemblance to the burial pit 3 and their
location in the same part of the Raupa site. There 1s also likely to have been an unexcavated
burial pit in Area ITI (Prickett 1990:109, 150).

Burial 3. In metre square B5 and extending ca, 180 mm into C5 was a near circular pit
1050 x 900 mm in plan at the surface. It was dug with vertical sides from the top of Layer
3 to a depth of ca. 1 m. At the bottom was some fragmentary wood which may have been
laid on top of the human remains.

Two human burials lay on top of one another with their heads both to the west, facing
south. Only enough of the remains were exposed to show that this was a burial before the
pit was refilled. Thus there are some things of which we cannot be sure. Red ochre was thick
in the loose soil about the remains and coating the surface of the bones themselves.

The two individuals were buried at one time on top of each other. The upper, and most
fully exposed of the two was a young adult, probably male. He lay on his right side with
knees drawn up to the chest and left arm lying diagonally over the arm and chest of the burial
beneath. The skull had moved slightly, probably with pressure from the settling of soil in
the pit, and was lying on its base rather than on its side, facing south. The age of the
individual may have been 25-30 years.

All that could be seen of the lower burial in the partial excavations was the head, left
arm, right shoulder, part of the pelvis and some fragmentary rib bones. Placed in the pit
first, the body appears to have been lying on its back; the head was turned slightly to the
south. The left arm was lying diagonally across the chest with the hand brought up beneath
the chin. The individual was female to judge from skull features and pelvis form. The teeth
were heavily worn and the left first molar at least had been lost during life; this suggests
a mature adult, possibly 30-40 years of age.

Further observations can be made regarding the teeth of the two individuals in Burial
3. Both of the first molars of the upper body showed extreme wear, greater than that of the
other teeth and with the distinctive slope of wear towards the cheek which is characteristic
of the so-called ‘fern root plane’. This pattern of wear is associated with a diet which
contains a lot of rough foods; it is often found in remains dating from the late pre-European
period (Houghton 1980;128-129). The loss of the first molar by the lower individual may
show the final result of this pattern of wear which is that the tooth becomes loose in its socket
and eventually falls out. There was no visible evidence in either individual of any other
dental disease such as decay, calculus or gum disease. This too is consistent with the
expected results of a rough diet.

Burial 4. At the south-east corner of Area XI, ca. 300-400 mm depth from the surface, was
a shallow scoop in which were placed the trussed remains of a young male. The skeleton
was complete except for the head and first neck vertebra. The remains were orientated
north-east/south-west with the head, were it still present, at the south-west end.

56 PRICKETT

The body was placed on its front with the legs drawn up tightly under the torso. The
left arm was at the side with the elbow bent very acutely to bring the hand up beneath the
shoulder. The right upper arm, shoulder blade and clavicle (collar bone) and the upper neck
vertebrae had also been moved and were lying on the back over the ribs. The right forearm
was lying diagonally under the body. Some of the lower vertebrae had also been disturbed
but all were present. The disappearance of the head and displacement of neck vertebrae,
upper right arm and shoulder and lower spine all appear to result from subsequent
disturbance of the grave. Part of the sacrum of another young individual lay over the right
side of the burial.

The pelvis shape and long and moderately robust limb bones suggest that the
individual was a male. In the absence of the head, and hence the dentition, the best estimate
of age is made from the general maturity of the bones which suggest a young person ca.
18-19 years of age. The height of the individual, who had completed growth, was calculated
atca. 1.78m(5 ft 11 inches), whichis taller than the average of 1.74 m derived froma sample
of 44 pre-European Maori males by Houghton et al. (1975:330).

A brief examination before reburial showed that this individual suffered at least one
period of illness or malnutrition during his childhood. The evidence for this is seen in the
lower legs where a small line of abnormal bone (Harris Line) was visible in the broken shaft
of the bone. This line indicated disruption and then recommencement of growth. There was
also evidence ofa hand injury sufficiently bad to infect the bone of the right hand. The injury
had occurred some time in the past and was completely healed, Features noted at the back
of the spine, the clavicle and on the bones of the knees indicated an active life. The form
of bones at the front of the ankle suggest habitual use of a squatting position. There was no
evidence as to the cause of death.

Burial 5. Area X was extended to the north-east to fully expose Burial 5 in metre squares
17-8 and J7-8. As with Burial 4 there was no evidence of a pit, but nor in this case was there
any sign of a prepared scoop or hole. The body lay on uneven ground with the knees higher
than the rest. Indications are that it was laid out on the ground and covered to a depth of
150-200 mm in an area of black soil and some midden waste, Apart from the head the body
was not disturbed in any way which argues against it being left lying uncovered,

The skeleton was complete apart from some foot bones, and was lying on its right side
with knees bent and slightly drawn up. The arms were lying at the front of the body with
the elbows bend. The right hand was beneath the head and the left hand close to it. The skull
had rotated slightly backwards and was crushed.

The pelvis which gives the most accurate information for determining sex, was too
badly damaged to be useful. The shape of the forehead, eye sockets and lower jaw, together
with the general slightness of the bones, suggested that the individual was female. The teeth,
including the stage of eruption of the third molars (wisdom teeth) and the immature state
of a number of bones, enable determination of age of death to within a narrow range of 16-
18 years. The height was calculated at 1.57 m which is consistent with the sex and
immaturity of the individual.

RAUPA 57

The teeth, almost all of them present, showed no evidence of decay and only slight
evidence of other dental problems such as calculus and gum disease. Most teeth showed
minimal wear. Exceptions were the upper and lower incisors which showed slightly greater
wear, and in particular the right upper central incisor which showed an unusual wear pattern.
Not only did this tooth show greater general wear but as well there was a shallow v-shaped
groove worn in the centre of the biting surface. This suggests that the front teeth, and this
one in particular were used more heavily than the rest, either in eating or possibly in the
preparation of fibre or some other manufacturing material.

Burial 6. Extending 400 mm into metre square A2 from the west was aca. 600 mm diameter
pit. This was not found during the 1987 work in adjacent Area III (metre square J4) although
some fragmentary wood close by may be related (see Prickett 1990:109).

The pit was cut from the surface of Layer 3 and must therefore post-date the main
occupation period of Raupa and predate the late 19th century mining upstream. In the soft
brown/grey fill ca. 250 mm below Layer 2B (see Fig. 10) were found a complete adult
human tibia, a large lump of ochre and charcoal. The pit was excavated only to ca. 750
mm depth - enough to show a marked reduction in diameter with depth. Nor was the
excavation extended beyond the Area VII baulk. Thus the full extent of human remains and
any associated material are not known.

Burial 7. In metre square C2 a 700 mm diameter pit was apparent at the surface of Layer
2 as an area strongly marked by charcoal and ash (see Fig. 17). On top of the fill at the north-
east corner of the pit was a massive water-rolled greywacke boulder. Around and partly
beneath it were three smaller stones which appear to have been placed to secure the larger
stone (Fig. 18). Thus at least some occupation of the Raupa site took place after the pit was
finally filled. Just below the fill surface was a piece of wood, possibly dressed, which
measured 170 x 50 x 5 mm. It was charred on one side. Ochre of a purple-red hue was
abundant throughout the soft fill.

Carefully placed in the pit, ca. 250 mm from the surface, were fragmentary human leg
and arm bones (humerus, tibia and femur) probably from a single individual. The pieces
were mostly the ends of bones together with small splinters from the shafts. They came from
a large robust individual, probably male. Signs of moderate arthritis at the ends of the leg
bones suggest a long and physically active life.

The bones found in burial pit 7 may be residue after removal of shafts for the making
of artefacts such as fishhooks, ornaments and bird spears. The femur, tibia and humerus are
the strongest and thickest parts of skeletal bone and typically were chosen for manufacturing
purposes. The cache of fragments here may simply represent disposal of the remains after
industrial use.

The pit has a total depth of 900 mm, a shelf half way down greatly reducing the floor
area. In the fill at the bottom of the pit was more bone and plentiful red ochre.

The evidence here is for a pit used at first for burial or the disposal of human remains,
which were subsequently exhumed to leave small pieces of bone in the ochre-rich soil at
the bottom. Later, fragmentary long bones were buried with some care at a shallow depth.

58 PRICKETT

Burial 8. A circular pit 800 x 700 mm in plan and 480 mm deep in Area XI, metre squares
E5-6 and F5-6, was originally dug from the surface of Layer 3. High in the fill were several
dog bones. From ca. 300 mm depth ochre was abundant in the soft material, but below was
only one human patella or kneecap. Comparison with other pits and the abundant ochre in
the fill strongly suggest there was a human burial here. It is possible the pit was re-excavated
at some time in the past and the bones removed for secondary interment or industrial use.

Burial 9, Extending from the west baulk of Area X in metre squares A9-10 was a 1.35 m
deep pit, ca. 900 mm in diameter at the surface of Layer 2 from which it was cut. At the
bottom of the pit were some fragments of human bone and abundant red ochre.

The shape of this pit, its proximity to other pits which were, or appear to have been,
dug for burials, and the presence of human bone and ochre combine to suggest that this was
a burial pit. It resembles burial pits 6, 7 and 8 in the presence of only a few bones, and pits
3, 7 and 8 in the presence of abundant ochre. As was the case with other burial pits the bone
at the bottom of the pit was not fully excavated and so is not described here.

Burial 10, In the 1987 excavation season a damp depression at the south end of the house
in Area III (Prickett 1990:109, 150) was not excavated. Experience in 1988 shows that this
probably marked the presence of a burial pit.

Adzes

Two adzes and several polished adze fragments were found in Areas VII, [X, X and
XI. Also, one adze was found between the excavations and the river bank. These finds bring
to eight the number of complete adzes or substantial pieces found in the 1987 and 1988
seasons.

A complete well-finished adze was found at the midden layer surface in Area IX, at
the boundary of metre squares K9 and K10 (Fig. 20). The raw material is a fine-grained
green greywacke with small black inclusions - sometimes called ‘chipwacke’ (Schofield
1967), The length is 118 mm, the blade width 48 mm and the depth 26 mm. The adze has
a rounded sub-triangular cross-section. The bevel and most of the front is polished, with
the sides, back and butt end of the front highly polished over hammer-dressing. The
undamaged blade clearly was not used after last being ground and polished.

In the metre square ES on the surface of Layer 3 was a second adze (Fig. 22). Again
it is made of fine-grained green greywacke and is of sub-triangular cross-section, although
in this case it is sharply defined at the angles. The bevel and much of the front is finished
to a high polish, with the remainder lightly polished over hammer-dressing. One side of the
adze back has not been polished at all but finished by fine hammer-dressing only, The adze
has been substantially damaged at both ends and only ca. 20-25% of the blade remains. It
measures 128 x 54 x 27 mm.

Outside the excavated area, 5.7 m east of the south-east corner of Area VII, a complete
adze was found in disturbed silt and mine tailings, left after this part of the Raupa site was
cleared by road grader prior to excavation (Fig. 21). The stone is green greywacke with
black inclusions. The length is 130 mm, blade width 43 mm and depth 24 mm. It is thus

RAUPA 59

Figs 20-27. Raupa adzes and patu muka fragments. 20. Adze, Area IX K9-10. 21. Adze,

surface find. 22. Adze, Area XI E5. 23. Adze blade, Area VI J5. 24. Patu muka handle,

Area VI CS, 25. Patu muka handle Area VI ES. 26. Patu muka handle end, Area
X E8. 27. End of patu muka, Area VII J2.

60 PRICKETT

the narrowest of all Raupa adzes for which measurements are possible. The cross-section
is rectangular and sharply defined, It is completely polished except for small areas at the
butt, and there is haft polish on the butt end. The blade shows signs of heavy wear.

Five polished adze fragments were found, all of green greywacke. The largest (metre
square F1, surface) was a 64 mm long piece of two sides of a rectangular cross-sectioned
adze, polished over hammer-dressing, Found in the disturbed surface silt in Area IX was
a 22 mm long fragment of a highly polished rectangular adze in very fine-grained
greywacke. Also highly polished was a 24 x 16 mm fragment of a sub-triangular adze from
square A3, surface layer. Another probably sub-triangular and highly polished adze is
represented by a tiny fragment only 14 x 9 mm from square L7, Layer 1. From the square
J2 surface came a 12 x 10 mm fragment of a polished adze in green greywacke.

Patu muka

In Area VII J2 was found the beating end of a massive flax beater made of pale grey
andesite (Fig. 27). Some original surface shows that the tool was fashioned from an
elongated water-rolled stone.

In the upper fill layer in Area X, metre square E8, was found the butt end ofa patumuka
made of coarse grey andesite (Fig. 26). This brings to six the flax pounders found during
the two years work at Raupa, four butts (handles), and two beating ends.

The Area X find is unusual in that it is very much larger and more circular than the
other three butts. The end is ca. 64 mm in diameter whereas the others are oval in shape
48 x 37 mm, 47 x 42 mm and 39 x 32 mm respectively, The well-formed knob reduces to
a slightly oval grip of ca. 49 x 53 mm. As with other patu muka recovered at Raupa the
particular form is well represented among Oruarangi material in the Auckland Museum.

Patu onewa

Part of the handle of a beautifully finished patu onewa (Fig. 28) was found at the west
side of metre square A8. In the 1987 excavation in adjacent Area III were found two blade
pieces of the same patu so that it was anticipated that more would be found in the next
season's work. Not all of it was found however and just where the remainder came to rest
is not known. The raw material is grey/green Waiheke Group greywacke.

In the 1987 excavation report it was suggested that the patu may have been heat
fractured during the fire which burned down the large house (Prickett 1990:118), One of
the blade pieces showed signs of heat damage. The missing pieces, clearly some distance
away, add weight to an alternative explanation, which is that the weapon was deliberately
destroyed or, more likely, broken during fighting. The large blade piece has been in a fire
and there is no crushed edge nor were small fragments found which would positively tell
of the destructive blow; nonetheless, breakage during conflict must be a likely explanation.

Nephrite

In Areas VII, [X-XI were found seven artefacts or worked fragments of nephrite. A
range of artefact types is included.

RAUPA 61

Fig. 28. Patu onewa found in three pieces: in 1987 (Area III; see Prickett 1990:118,
146); and 1988 (Area X A8), the upper (handle) fragment.

62 PRICKETT

Figs 29-35. Nephrite artefacts from Raupa. 29, Pendant, Area XI A5, 30. Pendant

fragment, Area VII 11, 31. Pendant fragment, Area X H9, 32, Chisel, Area X C8. 33.

Chisel butt fragment, Area VII A3. 34. Chisel butt fragment, Area VI J5. 35. Butt and
blade fragments from single chisel, Area VII G2.

The most important item, at least partly because it is still complete among all the
fragments is a small pounamu pendant (Fig. 29). It was found in square AS close to the
western side of the excavation in the upper occupation level. The pendant is triangular in
shape, 44 x 21 mm and 4 mm in maximum thickness, highly polished front and back. The
rear is more or less flat but the front is attractively curved to show off the high quality
lustrous emerald green stone to best advantage. Two suspension holes show this was a
necklace unit and not an ear pendant, as indeed the shape suggests. One hole has been
grooved front and back before drilling while the other has been drilled right through.

Among surface material in metre square H9 was found a second broken item, in all
probability being fashioned as a pendant (Fig. 31). It measures 37 x 22 x 7 mm, has been
extensively polished on both sides and at one squared-off end, and has been grooved front
and back where a break has occurred. The very broad groove suggests that it was part of
the fashioning of a once larger pendant and not necessarily for deliberate breaking at this
point, a suggestion endorsed by the very rough break which has occurred. If this was the
case then an unknown amount of the pendant has been lost from below the groove; it is
possible that the item was to have been an adze-shaped pendant, with a decorative groove
such as that illustrated by Skinner (1974:91). The suspension hole is drilled from one side
only probably to present the other side to the viewer when worn. The mid-green nephrite
is of comparatively poor quality.

A third pendant is represented by a drilled fragment recovered from surface layers in
metre square I (Fig. 30). This was an ear pendant of the kuru (see Law 1980), or possibly
kapeu (Te Rangi Hiroa 1949:286), form. The broken out suspension hole is worn smooth
indicating much use. The stone is high quality dark green.

RAUPA 63

The remaining four pieces of nephrite probably represent three chisels. Outstanding
is acomplete example from metre square C8 which is only 22 mm in length and 9 mm across
the blade (Fig. 32). The maximum depth is 5 mm. This tiny chisel is highly polished on all
four sharply rectangular sides. The raw material is a dark green nephrite with black and
white veins. On the bevel and back of the blade can be seen use marks, those on the bevel
interestingly being 10-15 degrees from the length of the chisel showing that the chisel was
habitually driven into the wood at that angle, the right corner first.

Two fragments from metre square G2 probably belong to the same chisel (Fig. 35).
The blade end is 22 mm long with a flat front and rear and rounded sides 10 x 6 mm. The
butt fragment is only 10 x 8 mm and appears to have been broken both ends, The raw
material of both items is an identical high quality mid-green nephrite and both pieces are
highly polished to the same appearance. The complete item may have been ca, 50 mm in
length.

Very different nephrite is used for the last piece, a chisel butt found in square A3,
among surface material (Fig. 33), The stone is an opaque olive green jade of high quality.
The length is 23 mm and the cross-section only 10x 7 mm. There is a high degree of overall
polish and some haft polish on the butt. Again the complete chisel may have been ca. 50
mim in length.

Bone artefacts

Bone artefacts from this part of the Raupa site include several fragmentary tattoo
chisels, bird spear and needle fragments and pieces of sawn and ground bone. The 24 mm
long piece of bird spear was found among midden waste in metre square K9 near the south-
east corner of the excavated area (Fig. 36). It has two large projecting barbs, unlike the bird
spear fragment with small barbs found in 1987 (Prickett 1990:147) but similar to a complete
specimen from Oruarangi (Fisher 1934:Fig. 74), The raw material is mammal bone. From
square G3, Layer 3, came a bone needle fragment 18 mm in length and 3-4 mm diameter
(Fig. 37).

The remains of four tattoo chisels were found together in metre square F5. These
fragile items were damaged during excavation, although three have been substantially re-
assembled from pieces (Figs 39-42). Two nearly complete chisels measure 46 x 9 mm (Fig.
39), and 48 x 10 mm (Fig. 40) respectively. Another 42 x 9 mm item shows only a small
part of grooves scored to form teeth at the working end (Fig. 41), The complete item might
have been 8-10 mm longer. Five further fragments include one lashing end, 16x 6 mm (Fig.
42), notched for attachment, and two grooved fragments. It is not clear ifthese pieces belong
to one or more chisels. Of interest is the fact that all four Raupa chisels belong to Fisher’s
(1934:284-285) notched type, like only seven of 75 complete items from Oruarangi. One
more Raupa chisel is represented by the plain butt found in 1987 (Prickett 1990:Fig. 49),
belonging to Fisher’s Type C which constitutes 54 of the Oruarangi group,

There are four pieces of sawn and ground bone, all from Area 1X. A 36 mm long split
half of bird bone (Fig. 38) assigned only to surface silt, has been sawn al both ends, possibly
to make a toggle. In the midden layer, metre square M10, was found a piece of mammal
long bone 70 mm in length, 12-20 mm wide and 6 mm deep which had been ground down

64 PRICKETT

both sides (Fig. 44). A cut has been started 2-4 mm in from one long side on the outer surface
of the bone, perhaps in preparation of a bone needle. Also from the Layer 3 midden (metre
square L10) was a fragment of similar bone sawn and snapped at both ends, polished on the
other four sides, and scored both ways on the smooth outer surface of the bone (Fig. 45).
It measures only 18 x 12 x 8 mm and must surely be waste. The last fragment was found
above the midden layer in square M6 (Fig. 43). It measures only 18 x 8 x 3 mm with two
sides and a straight edge fashioned by polishing and the remains of three shallow parallel
grooves on one side. It is likely this fragment results from a break during manufacture and
was never part of a finished item.

Figs 36-45. Bone artefacts from Raupa. 36. Bird spear fragment, Area IX K9. 37. Needle

fragment, Area VII G3. 38. Sawn and split bird bone (toggle?), Area IX. 39-42. Tattoo

chisel fragments, Area XI F5. 43. Sawn fragment, Area IX M6. 44. Ground
mammal bone, Area IX M10. 45. Sawn fragment, Area IX L10.

Pumice

From the Layer | surface, metre square G1, came a cone-shaped piece of pumice 22
mm high and 22 x 20 mm across the flat base (Fig. 46). It is faceted all around the cone
through rubbing, as well as being polished flat at the base. It is possible this item was formed
in polishing or burnishing wood or bone artefacts; more likely, however, it was itself the
object of manufacture, the only suggestion coming readily to mind being that of a gourd
stopper.

Part of a pumice bowl was found in two pieces (which do not fit together) in square
E6, Layer 1. What remains suggests a bowl ca. 55 mm in external diameter with sides
slightly squared off, and 45-50 mm in height (Fig. 47). Internal measurements are ca. 25

RAUPA 65

mm diameter and 28 mm depth, almost identical to that of a bow! found nearby in Area [V
in 1987 (Prickett 1990:129, 147). A groove around the rim may have been intended to seat
a top to the container. From squares E-D2 came a natural pumice cobble ca. 100 x 85 x 55
mm in size with 6-8 mm deep depressions in each side ca. 30 mm in diameter (Fig. 59).
This item may have been tried out for bow] manufacture, the raw material in all probability
proving too coarse and angular. Four fragments of a bowl which may have been 45-50 mm
in internal diameter were found in metre square B4.

An unusual raw material was a piece of finely banded obsidian pumice from Mayor
Island from metre square HI. A natural slab ca. 100 x 55 x 27 mm has evidence of use as
a hoanga on one large surface. From loose surface material in square D3 came two fitting
fragments of more usual pumice ca. 60 x 45 x 25 mm in size. A flat surface may possibly
have been the result of use for grinding. Other pumice fragments with indications of use
in grinding or polishing came from squares H2, F2, DI and E5.

Figs 46-47. Pumice artefacts. 46. Cone-shaped piece (gourd stopper?) Area VII G1. 47.
Fragments of pumice bowl (broken during excavation), Area XI E6. Another small rim
piece is not shown.

Drillpoint

From the disturbed surface of metre square D3 came a fine-grained purple red chert
drillpoint, unusual in having all three edges ground close to the point.

Hoanga

In Areas VII, [IX-XI were found six angular hoanga fragments which had been used
for grinding polished stone tools. Among loose material on the surface of square Fl was
amuch used fine-grained siliceous quartz sandstone hoanga displaying five highly polished
facets (Fig. 52). One of the facets was ground into a dish shape. Elsewhere two sharp edges
have been grooved through use as a cutter or saw. The stone measures ca. 60 x 45 x 40 mm
and appears substantially complete. A similar hoanga from nearby square JI also found in
disturbed surface material was again used on five surfaces (Fig. 53). The raw material is
a coarse quartz sandstone. Three of the grinding surfaces are strongly dished. On one of
them are traces of kokowai and a 30 mm long groove, the result of sharpening‘a point or
blade.

66 PRICKETT

Other hoanga items were only fragmentary. A 98 x 78 mm natural flat surface of an
andesite block has been used for kokowai grinding (Fig. 51). Once again it was recovered
from among loose surface material, in square ES. Another natural slab of andesite 25 mm
through has been ground very smooth on one side where striations can be seen under
kokowai stains (Fig. 49). Kokowai can also be seen in two other places on the stone. It comes
from among surface debris in F1. A piece of identical material from square D3 can be fitted
to the F1 stone to make a total polishing surface 120 x 80 mm. The break is not fresh. Also
from square D3 among surface material is a fragment of similar slab andesite 30 mm deep
(Fig. 50). One surface may have been ground.

For cultural reasons, these images have been removed.
Please contact Auckland Museum for more information.

Figs 48-53. Hoanga. 48. Andesite slab, Area VI A-B6. 49. Fragment of andesite slab,

Area VII F1. 50. Fragment of andesite slab, Area VII D3. 51. Fragment of andesite slab,

Area XI E5. 52. Fine-grained sandstone, Area VII Fl. 53. Coarse-grained
sandstone, Area VII J1.

RAUPA 67

Kokowai

As in other parts of the Raupa site ochre was abundant throughout Areas VII, LX-X1,
It was found as tiny fragments of soft processed kokowai and also as occasional more
substantial pieces of raw material. A piece of raw material weighing 27.4 g from among
surface silt in Area LX shows both soft brick red material and the brittle magenta coloured
volcanic rock from which it has weathered.

Several pieces of kokowai show evidence of having been worked. From surface
material, metre square C2, came a 16.9 g piece of soft brick red raw material. A polished
scoop on one side indicates either direct use of the dry red colouring on an object or the
grinding off of fine powder for mixing into paint, commonly with shark oil. From square
L7 came a similar piece only 4.4 g, which also had been polished into a slight scoop and
with a small area of polish also apparent on another side, A fourth piece, from square E2,
took the form of the end ofa thin ‘stick’ of raw material, weighing 11.9 g, polished through
use except where it has been snapped off the originally larger piece. The raw material is
high quality brick red haematite. Also from E2 were recovered four fragments of total
weight 5.0 g, broken during excavation from a piece of soft, very fine, deep red haematite
no more than 20 mm in length. One side of the piece was dished from use, while other
surfaces showed evidence of rubbing, A square G9 fragment only 20 mm in its greatest
dimension is grooved on one side either through use as a tool or as the object of attention.

Rounded water-rolled stones stained with ochre have clearly been used for grinding
kokowai (Figs 54-58). The largest of these was a 150 x 120 x 60 mm (Fig. 57), and the
smallest, a cylindrical pebble ca. 70 mm in length and 32 mm diameter (Fig. 55). The latter
was smeared with kokowai down the sides, while the two ends showed the chatter marks
of use as a knapping stone; the raw material is greywacke. Another stone is smeared with
both kokowai and black pigment.

From metre square H10, Level I, came a very large piece of yellow ochre raw material
weighing almost 5 kg. There is no sign of it having been used, but clearly it has been brought
to the site for use, probably from somewhere in the adjacent ranges as almost certainly the
red haematite raw material will have been as well,

Obsidian

A total of 1440 pieces of obsidian weighing 7483.3 g were recovered from Areas VII,
IX, X and XI, Material was assigned during excavation to “surface”, “Layer I”, “Layer 2”
and “Layer 3”. Surface, Layer I and 2 obsidian all belongs to the second occupation period,
contemporary with the nearby house. The few pieces recorded as coming from Layer 3 total
less than 20 g, and for the purposes of analysis are included here with Level II material, to
which they may in fact belong. Obsidian distribution is given in Table 3.

Obsidian density varied greatly throughout the 99 m? excavated in this part of the
Raupa site, from 146.1 g per m? in Area XI, 94,8 g per m? in VII, and 34.1 g per m? in X,
to only 7.4 g perm’ in Area [X, From only 20 closely grouped metre squares in Areas VII
and XI came 4552.4 g, more than 60% of the total. Twenty-eight percent came from just

68 PRICKETT

For cultural reasons, these images have been removed.
Please contact Auckland Museum for more information.

Figs 54-59. Stones used in kokowai grinding, and pumice piece. 54-58. Kokowai

grinding stones. 54. Area XI E4. 55. Area XI D5. 56. Area XI F5. 57. Area XI D4. 58.

Not accurately located within Areas VII, [X-XI. 59. Worked piece of pumice, Area VII
D-E2.

RAUPA 69

6 m*. These two figures represent 227.6 g per m* and 350.4 g per m? respectively. In
contrast, at the south-east of the excavated area in Areas 1X and X just 304.3 g were
recovered from 30 m?’, an average of little more than 10 g per m?.

Some very large pieces include a 168.1 g flake of green obsidian with possible use
damage from square C4. From D6 came a 139.7 g core of green flow-banded material. Other
pieces of more than 100 g were a 117.1 g flake from C3, and two shatter pieces from F4,
108.4 and 101.8 g respectively. Flakes of more than 80 g were found in squares Al, D1,
and ES and 6. All these large pieces come from that part of Areas VII and XI where the
concentration of obsidian is greatest.

It is tempting to relate the large quantity of obsidian pieces in the area bounded by
squares C1-5 and F1-5 to the several massive water-rolled boulders found in this part of the
site. Sharp pieces of obsidian were essential in cutting flax, and the boulders may well have
been used as anvils for pounding scraped flax to make up the fine muka fibre used in cloak
making. It is not impossible that the burnt bowl in this area also had been used in the
manufacturing process, for soaking or dying the fibre. Consideration of the overall obsidian
distribution from the 1987 and 1988 excavations is included in the discussion section below.

As in other parts of the Raupa site most obsidian pieces are very small, Overall, 40.8%
are less than | g with another 35.3% of 1-5 g. Throughout the four areas (VII, IX, X and
XI) obsidian pieces of 0-5 g consistently made up 75-79% of the total. Ninety pieces of more
than 20 g weight made up only 6.26% of the obsidian count but almost half (47.9%) of the
total weight.

Over 94% of all obsidian by weight was green in colour and almost certainly Mayor
Island in origin. Of the remainder, 55 pieces weighing 235.5 g (3.15%) were olive grey in
colour and identified as having come from the Waimata stream, Waihi, source (Moore and
Coster 1989). Forty-one pieces of 194.9 g (2.6% of total) were grey obsidian which may
have originated on the Coromandel Peninsula, on Great Barrier Island or, possibly, in the
central North Island. Grey and Waihi obsidians share with the collection as a whole the large
preponderance of very small pieces. Among grey items only one weighs more than 20 g
(25.3 g) and only eight others weigh more than 10 g. Three Waihi pieces are 43.1 g, 30.1,
25.0 g and 24.3 g respectively, with another two pieces more than 10 g.

Of some interest is the fact that ten pieces of scarce Waihi obsidian include areas of
original cortex, as do five pieces of pale grey material. The percentages are 18% ofall Waihi
pieces, and 12% of grey. In contrast only 12 pieces of green obsidian included cortex, that
is, less than 1%. Waihi and grey material was presumably collected as relatively small
water-rolled cobbles and pebbles. An example is a 24.3 g item from square A9 which
includes the greater part of a water-rolled stone of Waihi obsidian. Mayor Island material
on the other hand was, and still is, available as substantial pieces originating in large flows.

The 1440 pieces of obsidian can be roughly divided into flake pieces, shatter pieces
and cores. Approximately 60% of all pieces are struck flakes, 38.5% are shatter pieces and
the remainder cores or blocks of obsidian, Core pieces are mostly small remnants of once
larger pieces from which sharp-edged flakes have been struck. An exception well suited to
further flake productionisa 139.7 gcore or block of green flow-banded obsidian from metre

70 PRICKETT

Table 3. Summary of obsidian distribution in Areas VII, [X-XI; of a total of 1440 pieces
weighing 7483.3 g. Refer to Fig. 14 for metre square locations.

Sq. No. Wt (g) Sq. No. Wt (g)
Al 2 92.4 A2 l 4]
Bl 1 0.5 B2 8 24.7
il 7 36.8 Oy. 40 212.8
D1 13 200.2 D2 25 203.6
El 21 68.5 E2 RS. 116.6
Fl 33 111.8 F2 40 152.0
Gl 28 119.8 G2 30 47.2
H1 24 73.8 H2 vi 5.9
I] 34 83.1 je 22 94.1
Jl 13 535s J2 £3 75.8
A3 7 23.4 A4 1: 109.6
B3 10 49.0 B4 30 225.4
C3 PA| 358.1 C4 34 442.4
D3 29 135.8 D4 29 199.2
E3 17 64.1 E4 aD $9759
F3 b! i F4 28 408.3
G3 37 218.1
H3 30 97.0
13 13 41.6
J3 16 42.0
A5 ty 65.6 A6 9 id
BS 47 161.6 B6 9 40.4
CS 25 Sis C6 Le. 31.4
D5 75 i Via | D6 80 281.2
ES 30 254.4 E6 Li 24.1
F5 fl £230 F6 2 py
K6 l 3.4
L6 l pete
M6 v3 ae
A7 14 56.3 A8& 8 24.1
B7 14 89.0 B8& l 3.0
C7 10 32.8 C8 10 88.6
ibe] 12 Lae D8 9 29.8
E7 8 21,4. E8 13 329
F7 26 144.6 F8 19 34.5
G8 4 44.8
H8 2 the Pi
18 2 12.6
J8 19 53:6
K7 K8
es, ps 10.5 L8 3 8.9
M7 l ie! M8
A9 2 mae, Al0 6 28.7
B9 3 52 B10
C9 10 108.0 C10 6 32.1
D9 10 94.1 D10 12 48.1
E9 14 50.6 E10 7 24.9
F9 > 29.2 F10
G9 ii TZAl G10 ]

RAUPA 7]

Table 3 (cont.)

Sq. No. Wt (g) Sq. No. Wt (g)
19 4 42.6 110

J9 i4 42.1 J10 12 13.4
K9 2 2.1 K10 | 3,9

L9 2 0.7 L10 I 7.0
M9 ! 0.9 M10 3 31.4

square D6. From square F2 came a 50.7 g core remnant with battered edges, again of flow-
banded green material. Other core remnants weigh as little as 2.8 g.

Among flakes and shatter pieces approximately 30% has evidence of use in the form
of edge-damage. Most edge-damage shows where very small flakes have been taken off
sharp edges in the course of use in cutting or scraping tasks, A small number of pieces are
notched as if they have been used as spoke shaves. More pieces have battered or crushed
margins. One unusual 3.9 g green flake from square D4 has evidence of grinding on one
edge with some kokowai adhering. Interestingly, struck flakes and shatter pieces both have
edge-damage, showing that when a sharp edge was needed the users were not particular
about how this was achieved.

An interesting piece from square G9 is a 39.9 g core showing very clearly the method
of producing small flakes from the ca. 45 x 35 mm base. At the opposite side is poor quality
stone from the original flow margin. A 32.7 g piece from square A9 illustrates a very
different style of flake production, this time from the 32 mm long sides of a core with well
prepared platforms at each end. Yet another method of flake production can be seen on a
37.2 g piece from ES. This item shows flake production from all surfaces with striking
platforms used as they became available by the removal of flakes from adjacent facets.
Some edges are crushed from use for knapping or hammering after flake production
finished.

Flakes and shatter pieces alike have a range of shapes. Unusual is a broad 47.3 g flake
from square A&, ca. 90 x 50 mm and 15 mm deep at the water-rolled striking platform. The
long and well-shaped knife edge shows slight damage through use. In contrast is a 9.5 g
blade-like flake 58 mm in length and 24 mm maximum width (square B8). Considerable
edge damage can be seen along a concave side but none on the other, convex, side.

Chert

Throughout the 99 m? of Areas VII, IX, X and XI were found 529 pieces of imported
stone classified as chert, weighing a total of 8699.8 g. The weight represents a high density
of chert compared with other parts of the Raupa site. In the adjacent Areas III and IV, for
example, were 33 g and 69.5 g per square metre respectively whereas from the area now
under discussion came 84.9 g of chert per square metre. From Areas VII and Xl alone came
more than 100 g per square metre. Details of chert distribution are given in Table 4.

Chert, like obsidian, is a fine-grained rock from which sharp-edged knives were struck.
The predominant material, as elsewhere in the Raupa site, is white or cream coloured with
pink, red, yellow and sometimes grey or brown, Cortex is usually dark red or brown. This
chert may be coarse or fine-grained and varies greatly in its suitability for flake production,

72 PRICKETT

Table 4. Summary of chert distribution in Areas VII, [X-XI; ofa total of 496 pieces weighing
8339.5 g. Refer to Fig. 14 for metre square locations.

Sq. No. Wt (g) Sq. No. Wt (g)
Al A2
Bl l 206.0 B2
ce 2 1.4 Oe, > 29.9
D1 5 9.4 D2 16 370.0
El 14 140.6 E2 5 3.9
Fl 3 43.1 F2 13 262.9
Gl 14 115.6 G2 10 372.9
HA 4 32.8 | Wa 3 29.9
I] 8 38.3 12 l yas
Jl 3 26.2 bes 8 8.3
A3 3 0.7 A4 i 1.4
B3 y. 5.3 B4 5 6.1
C3 6 164.2 C4 5 70.2
D3 12 617.0 D4 163 322.8
E3 5 502 E4 es 99.7
F3 v4 555 F4 LS 237.0
G3 10 170.3
H3 8 101.3
is 5 198.3
JS 4 11.0
AS 5 36.1 A6 8 8.5
B5 7 55.9 B6 4 140.4
C5 C6 9 39.2
D5 9 i Ae D6 10 53.9
E5 9 487.3 E6 4 1.8
F5 S 93.2 F6 l 6.2
K6 l 1.9
L6
M6 Za 5.9
A7 9 131.4 A8 7 702.5
B7 6 364.5 B8 6 37.8
C7 10 258.5 C8 10 34.4
D7 13 56.6 D8 18 68.1
E7 2 oi. E& 2 4.8
F7 10 89.4 F8 ] 12.3
G8 l hs
H8 6 43.3
I8 | 0.9
J8 3 L1e3
K7 = PAY K8 5 38.5
| (ors L8 1 1.8
M7 M8&
A9 14 179.0 Al0 4 mt
B9 6 a2i2 B10 6 19.0
C9 5 95.8 C10 8 8.5
D9 5 68.1 D10 4 140.4
E9 9 358.0 E10 9 39.2
F9 2 py F10 10 53.9
G9 2 4.8 G10 4 1.8
H9 3 8.6 H10 l 6.2

RAUPA 73

Table 4 (cont.)

Sq. No. Wt (g) Sq. No. Wt (g)
19 “i 57.4 110 2 3.5
J9 J10 10 89.4
K9 3 74.6 K10 ] 2.5
L9 L10 ] 4.3
M9 l 3.6 M10

Some very fine-grained chalcedonic material was commonly white or brown. Rare pieces
of jasper were mostly small.

Some pieces of chert show cortex that has been severely battered by water-rolling.
Other cortex shows no sign of having been water-rolled, the material presumably having
originated in outcrops. It is assumed that all the chert came from the nearby Coromandel
Ranges where it is common in variety geological contexts.

Like obsidian, the major concentration is in the area confined by squares C1-5 and
G1-5, with a minor concentration at the south and west of Areas X and XI. Only 20 m? of
the 99 m? total have more than 76% of the total. Like obsidian, the very marked
concentration of chert in Areas VII and XI suggests this was a working area. In this case
there is evidence of considerable stone working, including the preparation of chert cores
and flake production. Chert flakes may also have been used in fibre preparation, as was
suggested for the obsidian concentration; again this may be associated with the massive
water-rolled boulders here which arguably were brought in as anvils for use in muka
production.

The concentration of chert weight in only a few metre squares depends on the presence
of a small number of very large pieces. Thus only twenty pieces of more than 100 g each
make up 55% of the total weight. One of these pieces, from metre square 12, is of a fine-
grained cream and pale brown chert weighing 775.3 g (Fig. 60). The stone has had almost
all cortex removed and takes the form of a half-sphere core from which flakes have already
been removed and which is well prepared for further flake production. Another large piece
weighing 415.2 g (square D3) is of coarser cream material with some red-brown cortex (Fig.
61). Again it has been extensively used for flake production. An unusual piece of chert from
square G2 showed signs of extensive use as a hammerstone (Fig. 71). The raw material is
cream with orange veins. The piece weighs 1 18.3 g and measures ca. 40 x 40 x 40 mm. Other
large pieces of chert are illustrated in Figs 62-70 and described in Table S.

Of the 529 chert pieces 54% are shatter pieces, almost all unused waste, and ca. 42%
are struck flakes. Edge-damage as a result of use could be seen on ca. 10% of pieces. The
remaining 4% are blocks or chunks of rock or cores modified to greater or lesser extent for
flake production.

74 PRICKETT

For cultural reasons, these images have been removed.
Please contact Auckland Museum for more information.

Figs 60-71. Major chert pieces. 60. Area VII 12. 61. Area VII D3. 62. Area X. 63. Area
X A&. 64. Area VII C3. 65. Area VII 13. 66. Area XI C7. 67. Area XI B7. 68. Area VI
G3. 69. Area VII B1. 70. Area X A8. 71. Hammerstone, Area VII G2.

RAUPA 75

Other stone

Other stone imported to the Raupa site includes greywacke, petrified wood, quartz and
a wide range of hangi stone raw materials, most commonly greywacke and andesite but
including also basalt, red (altered) rhyolite, indurated rhyolite and indurated mudstone.
Throughout the site hangi stones made up the overwhelming bulk of imported stone
material.

Petrified wood is of some interest because of the use of the material for lure shanks
where longitudinal strength was required. Twenty pieces were found in the 99 m?, of which
two made up 67% of the total 369.6 g weight. Three or four pieces are long enough for use
as lure shanks although there is no sign of their having been worked on. A 45 x 20 x 6mm
piece (square 19), broken offa larger block, has evidence of grinding use on a narrow curved
surface,

A 958 g piece of water-rolled greywacke cobble from square F3 has had two large
flakes removed from one end. The two concave edges formed as a result show considerable
edge-damage possibly through use in the manner of a spoke-shave.

Andesite occurred as fragmentary hangi stones, and as flaggy slabs, occasionally used
as hoanga which have already been discussed. It is likely that all the andesite slabs at Raupa,
whether or not they have evidence of use, were imported for use as hoanga. An unused piece
from square BS shows signs of having been deliberately split to present a flat grinding
surface.

The largest piece of quartz, from square K9, weighs 70.2 g and takes the form of an
angular flake with a notch ca. 25 mm wide and 10 mm deep on one edge, showing signs
of intensive use by its very crushed appearance. Other small pieces of quartz were imported
possibly as an abrasive to aid the process of grinding stone tools.

Kauri gum

Four lumps of kauri gum were recovered. All of the 28.6 g total, excepting one tiny
fragment, came from squares A2 and C2, to suggest that the distribution was not entirely
accidental. The three pieces are of good quality and it is possible they were gathered as a
raw material for use such as in the manufacture of tattoo pigment.

Material of European origin

Some fragmentary items of European origin came mostly from among disturbed rock
flour and silt. Included were of fragments of a plain white ceramic bowl (square H10), a
31 mm long copper nail (G1), and four green bottle glass pieces (El, J2, D3 and G3). All
of this material belongs to the late 19th or early 20th century, that is, after abandonment
of the Raupa settlement. More interesting area clay pipe bowl fragment, without decoration
or mark, from square B8 Layer 2 and a piece of totally rusted iron of uncertain shape from
a posthole (see Fig. 16) in squares D1-2. Both these items must be assumed to date from
the Maori occupation of Raupa.

76 PRICKETT

Table 5. Chert pieces of more than 100 g weight from Areas VII, IX, X and XI.

Metre
Square

Wt(g)

206.0

188.8

176.2

137.1

118.3

775.3

156.5

415.2

119.7
white chert

149.8

Rock characteristics

Fine-grained but poor
flaking quality;
red/brown with
extensive yellow/
off-white cortex

Fine-grained off-white
with grey/brown cortex

Uniform coarse-grained,
dun-coloured stone with
dark grey cortex

Cream and some quartz
banding; comparatively
coarse

Variable quality white
and cream chert with
orange veins

Fine-grained cream/
brown chert of good
quality throughout

Orange/red/white fine
-grained chert; some
quartz

Coarse-grained cream
material; red-brown
cortex

Uniform fine-grained
well-prepared platform;

Chalcedony quality;
translucent white/
ochre/red/black

Evidence of use

Some flakes struck
off intransigent
raw material

Angular struck piece
with evidence of
previous flaked
surfaces

Flakes have been
struck off this piece
which may itself have
been a flake

Extensively flaked all
around; step fractures

Hammerstone with
surface battered
through use

Well prepared for
flake production -
domed core with many
flakes already removed

Large angular flake
with step fractures
and evidence of
previous flaking on
outer surface

Core extensively used
for flake production
Flake struck from
possible edge damage
Battered edges of

angular flake show
considerable use

Fig.

Table 5 (cont.)

Metre
Square

Wt(g)

191.5

148.7

276.2

127.8

328.8

158.7

340.9

222.6

112.0

316.7

Rock characteristics

Fine-grained high-
quality off-white chert
with pink tinge in
places and orange/red
cortex

High quality chalcedonic
chert; cream/yellow/grey
with orange/red cortex

Coarse off-white/
cream sinter

White translucent
chalcedony water-rolled
pebble; yellow cortex

High quality chalcedonic
chert; cream/yellow/grey
with orange/red cortex

High quality very fine-
grained black chert with
orange/red water-rolled
cortex

Off-white/yellow/red with

deep red cortex; mostly
fine-grained but with
inclusions of coarse
crystalline sand

Variable rock coarse to
chalcedonic quality;
largely grey

Material similar to above

but more uniform in colour

and quality

Part of large water-
rolled fine-grained
stone; deeply
weathered red/brown
cortex and yellow/
blue/grey interior

Evidence of use

Large flake, edge-
damaged through use

Large well-formed
flake; some edge-
damage through use

Waste block, possibly
fire shattered

Flakes removed from
one end

Massive well-formed
flake, some edge-
damage through use

Large well-formed
flake, with good edge
apparently unused

Core; flakes removed
from several surfaces

Large shatter flake
with previously
flaked outer surfaces;
no sign of use

Struck flake with
possible use-wear
at one sharp edge

May have been struck
to form a core but
poor fragmented
quality has ruled out
further use

Fig.

RAUPA 77

78 PRICKETT

AREA VIII

Area VIII was 23 m north of Area V (excavated in the 1987 season), and more than
40 m from the northernmost part (Area IV) of the extensive central excavation of the 1987-
88 Raupa programme (Fig. 2). The location was chosen to be well within the innermost of
the three defensive ditches identified by Phillips (1986:99), thus it was hoped that
something would be learned of the use of this area, thought to be immediately behind the
main defensive lines of the pa. A 5 x 5 m square was opened up, later extended by | m?
at the east side (26 m7’).

Although 15-20 m south of the inner ditch identified by Phillips (1986:99-100),
archaeological evidence in the Area VIII excavation square proved to be of further
defensive works and not of living space as had been anticipated. Several massive stockade
postholes, some with post butts still in them, were found in the north-east part of the square.
To confirm and further investigate these defences two further trenches were dug by machine
east and west of the excavated square (Fig. 72).

The 5 x 5 m square

Stratigraphy was complex. In the initial 5 x 5 m square it comprised a series of thin
layers dating from before and after the digging of massive holes for the stockade posts.
Stratigraphic layers illustrated in Fig. 73 are as follows.

Layer 1. Grey sandy clay with rare charcoal.

Layer 2. Pale grey clay.

Layer 3. Grey to black silty clay with plentiful charcoal and shell.

Layer 4. As Layer 3 and including also redeposited brown clay lumps.

Layer 5. Predominantly grey clay with charcoal and yellow clay lumps.

Layer 5a. Predominantly yellow/brown clay with charcoal and some lenses of black silt
loam.

Layer 6. Pale grey clay posthole fill with rare charcoal.

Layer 6a. Mixed yellow/brown and grey clay posthole fill with plentiful charcoal and
some shell.

Layer 7. Black clay, full of charcoal.

Layer 7a. As Layer 7 but not so black and includes shell.

Layer 7b. Grey clay, rare charcoal.

Layer 8. Loose black soil with shell - mostly pipi, rare cockle (Austrovenus
stutchburyi). ;

Layer 9. Very black, charcoal rich silt loam with rare shell.

Layer 10. Loose brown sand.

Layer 11. Yellow-brown sandy silt; natural base of site.

The section drawing (Fig. 73) and photographs (Figs 75 and 76) show something of
the history of this part of the Raupa site. Two major occupation periods can be defined, The
palisade postholes have been cut through earlier material consisting of soil rich in charcoal
and with some crushed shell (Layers 7, 8 and 9). There were no ovens or even oven stones
here to suggest the charcoal and shell came from cooking operations in the vicinity, and
the complete lack of any postholes other than those of the later stockade gives no further

RAUPA 79

West trench

East trench

a 1984 survey peg

co)
a
—

co)

O Palisade post

-— Approximate line of defensive trench N

— — Approximate line of defensive ditch
(see Phillips 1986:99)

Fig. 72. Plan of Area VIII showing location of two machine dug trenches and Phillips’
1984 trench.

80 PRICKETT

N 1 2 DATUM a 5 iS)

Posthole

350mm diameter, 650mm deep
350, not excavated

500 x 600, 1500

450, 1700

400, 850

530, 800

500 x 400, 1100

Gs Mm & c a I

Figs 73-74. Area VIII plan and section drawing. 73. East section. 74. Plan.

RAUPA 81

For cultural reasons, these images have been removed.
Please contact Auckland Museum for more information.

Fig. 75. Area VIII, east section and one metre extension.

Fig. 76. Area VIII, view from west showing palisade line cut through earlier black,
charcoal rich, soil.

82 PRICKETT

clues as to what this part of the site was originally used for. The thick layer of black cultural
soil may, in fact, be the result of deliberate dumping here to level up the area prior to
construction of the defences.

Figs 73 and 75 show clearly the result of cutting through the earlier black occupation
layers to put the palisade line in place. Typically the front face of the hole (facing out from
the pa) has been cut vertically, while the rear face is stepped, probably to make easier both
the digging of the hole and the raising of the massive post which was to occupy it. The
vertical front face of the holes would strongly resist any attempt by an attacking force to
pull down the palisades.

Three of the seven postholes in Area VIII had decayed post butts still in place (see Fig.
77): two were of totara and one of kauri. The measurements are given as Fig. 74 where it
will be seen that six of the seven are between 800 mm and 1700 deep. These large posts
almost certainly made up part of a stockade. The seventh post (A) was only set 350 mm into
the ground and would not have been strong enough for a defensive post. The | m extension
(square F2) was excavated to confirm the close set stockade line here. Further confirmation
of the defences was sought in two trenches east and west of the main Area VIII excavation.

The trenches

The two trenches were dug by machine to a depth of ca. 2 m. Eastwards a 15 m long
trench was dug on a north-south line between the main square and the river, and to the west
was dug an |1 m trench.

The east trench (east) section is shown in Fig. 78. The metre squares are numbered
north to south. The stratigraphy is as follows.

Layer 1. Banded yellow rock flour, the result of late 19th and early 20th century mining
operations upstream.

Layer 2. Brown and grey mottled sandy clay with rare small pieces of charcoal.

Layer 2a. Clay with charcoal and abundant shell.

Layer 3. Light grey clay with rare charcoal lumps.

Layer 4. Blue grey sandy clay, with charcoal comparatively plentiful in squares 5-8 and
concentrated lens of charcoal in squares 4-5 (as shown),

Layer 5. Brown sandy clay.

Layer 6. Grey clay with plentiful fine charcoal.

Layer 7. Brown sand.

Layer 8. Black clay with dense charcoal, shell and oven stones.

Layer 8a. As for Layer 8 but without shell.

Layer 8b. As for Layer 8a and also including patches of light grey and brown clay.

Layer 9. Mottled clay with charcoal,

Layer 10. Light blue-grey clay - natural base of site.

Two outstanding features are shown in the section drawing: the first is a group of
postholes near the south end of the trench, and the second a broad and shallow trench at the
north end (see also Fig. 72). Two massive postholes in squares 13 and 14 are ca. 2 m and
1.3 m deep respectively. Very decayed and fragmentary wood from the southerly posthole
was identified as totara. Three metres north of the deeper posthole is a small angled posthole
filled with charcoal.

RAUPA 83

For cultural reasons, this image has been removed.
Please contact Auckland Museum for more information.

Fig. 77. Totara post butt from Area VIII palisade posthole D (see Fig. 74). Note that the
bottom of the post (right) has been relatively well preserved within the watertable. Scale
intervals are 250 mm.

The trench to the north of the posts is approximately | m deep, and three or five metres
across depending upon how it is measured. It is not clear if it is artifical or natural in origin:
it does not have to be artificial to have been used for defensive purposes. The depth is similar
to that of the nearby ditch identified by Phillips in 1984 which was 4.4 m wide and only
1.2 m deep (see Phillips 1986:97). Also, the more sharply defined ditch in the west trench
(see Fig. 79) appears to be part of the same feature which may therefore be expected to have
extended all the way across the isthmus, as did the three ditches located by Phillips
(1986:Fig. 7).

In the western trench there was also a ditch some metres north of massive posts which
probably made up part of a stockade. Stratigraphy is depicted in Fig. 79. Layers are as
follows:

Layer 1. Rock flour and redeposited silt.

Layer 2. Predominantly yellow rock flour and some silt made up of many fine layers.
Layer 3. Mixed rock flour and clay.

Layer 4. Brown sandy clay with some red-brown clay and charcoal.

Layer 5. Thin shell layer lying directly on top of Layer 6.

Layer 6. Dense black charcoal layer with shell in places.

Layer 6a. Posthole fill: very dark grey silty clay with plentiful charcoal and some shell.

84 PRICKETT

Layer 7. Light grey clay mixed with some orange silty clay. Grades into Layer 7a.
Layer 7a. Homogeneous brown clay.

Layer 8. Greyclay and charcoal with rare shell fragments increasingly common to south.
Layer 8a. Similar to Layer 8 but with dense shell fragments.

Layer 9. As Layer 8 but shell abundant.

Layer 10. A mixture of light grey, light brown and mottled clay.

Layer 11. Brown sand.

Layer 12. Homogenous blue-grey clay tending browner with depth.

Layer 13. Clay and charcoal.

Layer 14. Light blue-grey clay - natural base of site.

The remains of two posts were found in the west trench (see Fig. 72). The southward
one can be seen in the section drawing, beneath Layer 6 (Fig. 79). Fragmentary timber was
identified as miro (Prumnopitys ferruginea). The other probably extended into Layer 7 in
the same way but the ditch digger removed the top part of the post hole before stratigraphical
relationships could exactly be determined. Some 5 m north of the northerly post the original
ground surfact dips ca. 1.25 m into a ditch which narrows to 700 mm wide at the base.
Although this is very different to the supposed defensive ditch in the eastern trench yet there
are similarities. Both are about the same distance from the line of posts and both are ca.1 m
in depth.

C14 Sample “’

YH:

Charcoal timber

eye e
Concentrated
charcoal posthole fill

Fig. 78. Area VIII east trench, east section.

RAUPA 85

Radiocarbon date

The source of charcoal sent for radiocarbon dating is shown in Fig. 78. The archaeological
location is: Raupa Area VIII east trench, Layer 6, metre square 4. Layer 6, with plentiful
fine charcoal, was situated at the base of the presumed defensive ditch which cuts the north
end of the east and west excavated trenches. The position of the layer shows that it is erosion
material which has probably fallen into the ditch from the south side at the beginning of
the process of infilling, to be quickly followed by bigger blocks of material. Thus the layer
and dating sample marks the first filling of the defensive ditch when its usefulness was past.
The species identified may reflect the vegetation that was cleared and burnt at this end of
the site at the time the defences were put in place.

S 1 10 9 8 7 6 N

Approximate

| position of |!

|
|Palisade post,

Fig. 79. Area VIII west trench, west section.

86 PRICKETT

The radiocarbon sample of 8.3 g was made up of charcoal identified as follows:

Coprosma sp. 8 identifications
Schefflera digitata 6 "
Leptospermum scoparium (manuka) | ‘
Pteridium aquilinum

var. esculentum (bracken) l "
Identifications were carried out by Rod Wallace, Anthropology Department, University of
Auckland, who commented that the material made a good sample for dating.

The radiocarbon analysis was carried out at the University of Waikato, Radiocarbon
Dating Laboratory. Results are as follows:

Sample number: Wk-2041

Carbon 14 depletion (D '*C) 1s expressed in 0/00 wrt 95% NBS oxalic acid:
-18.0 + 6.1 o/oo.

The isotopic fractionation correction (6 ''C) is expressed in 0/oo wrt PDB:
8 BC + 27.9 o/oo.

Radiocarbon age: modern.

The modern radiocarbon date suggests that this part of the Raupa defences dates from
the last occupation of the pa in the early 19th century. In 1893 the surveyor Kenny observed
only three defensive ditches at the north end of the site (Kelly 1946:208), these three
subsequently being identified by Phillips (1986:99). The fact that much of the fill in the
ditch now discovered comprised rock flour from upriver mining operations suggests that
had Kenny visited the site a few years earlier he would have seen four defensive ditches,
or even more; although why some ditches should be filled and others still visible is not clear.
In any event it must be argued from stratigraphic and radiocarbon evidence that the ditch,
and probably also the adjacent and parallel stockade uncovered in 1988, both date from the
last fortification of the Raupa site.

Hoanga

From square D1, close to stockade post B (see Fig. 74), came a rectangular 60 x 45
mm slab of andesite showing kokowai and the evidence of considerable use for grinding
on one flat surface (c.f. Figs 48-51). The natural slab is 19 mm thick.

Stone material

Very little stone material was recovered from Area VIII. In the fill of the buried ditch,
east trench, was found a 25.9 g flake of green obsidian with possible unifacial edge damage,
and from metre square D4 in the main excavation was found another piece, also green, of
2.3 g. One piece of the abundant chert of variable white, red and brown colour, weighing
21.0 g, was found in metre square AS.

RAUPA 87

DISCUSSION
The Raupa settlement

In the 1987 excavation report was included a discussion of the arrangement of the
Raupa settlement as understood from the archaeological work of 1983, 1984 and 1987 (see
Prickett 1990:77-79, 139-143). The 1988 season has thrown further light on the way in
which the people of Raupa made use of available space for structures and activities.

Fig. 80 summarises what is known from archaeological evidence of the arrangement
of the Raupa settlement in the late period of Maori occupation ca. 1820. It must be
remembered that large areas are unexplored so that by no means all of the complexity of
the settlement is shown. Nonetheless, it is interesting to see that the large Area III house
is a focus of activity areas. In front is a compacted area which probably reflects the
considerable use of this space. A different kind of use is indicated by the large quantity of
waste stone material, the ‘anvil’ stones and other artefactual material found immediately
east and north-east of the house. This was a working area, very likely for a wide range of
manufacturing and other tasks.

The location of this working area, between the house and its open area in front and the
cooking and refuse disposal behind, reflects the group activities which took place here.
Behind the house (Area VI) was an open area while further away (Area II, Area [X and
Trench A) are cooking areas with some dumping of waste. Cooking and waste disposal also
took place in front of the house, possibly in relation to another unexcavated house which
stood nearby.

The conclusion from 1987 work that ca. 1820 the Raupa settlement was diminished
in size from earlier occupation (Prickett 1990:148) received some confirmation from the
second season’s excavation. Immediately behind the Area III house (in Area VI) an earlier
cooking area was transformed into open space without evidence of any particular use. Area
[X, where there was a house in the early period, was later used for waste disposal from
nearby cooking ovens - also, in effect, open space.

Burials

The 1987 and 1988 excavations uncovered a variety of burial evidence at Raupa. There
was a double burial in a pit (Burial 3), two trussed burials (1, 4), a secondary burial (2),
extended burial (5) and four burial pits from which the skeletal remains were largely
missing (6, 7, 8, 9). Another probable pit burial was not excavated (10). The two seasons
excavations at Raupa uncovered six substantially complete human burials. Salient facts
concerning the buried individuals and burial practice are outlined in Table 6.

The burials almost certainly date from the first half of the 19th century, most likely
from both before and after abandonment of the Raupa settlement in the early 1820s. The
fact that none show any sign of Christian burial practice indicates a final date prior to
conversion of the Hauraki people in the middle years of the century.

88 PRICKETT

Defences wall

Waste disposal

Cooking

Open space V 6 rn ae,
90808 Ate

Cooking and

food refuse disposal

Bou] Cooking

Open space

Fig. 80. Plan showing the main activity areas of the later (ca. 1820) Raupa occupation.

RAUPA 89

Where it could be determined, several of the presumed burial pits can be dated to
before final abandonment of the settlement (see Fig. 81). Burial pit 7 was capped by Layer
3 on which there was a large anvil stone propped up by small stones which must have been
put there after final use of the pit. Pit 8 was sealed by material which included dog bone
pushed into an occupation surface, this showing that people lived here after excavation of
the pit and any interment which may have taken place. Pit 9 had two postholes dug into the
fill surface. The double burial (3) was also capped by a tramped occupation surface; as was
pit number 6 which was found in Area VII, but of which there was no sign in Area III
excavated as it was only to the house floor level.

Thus, some time before people stopped living at Raupa they buried their dead within
the settlement. Just when this took place is not clear: it may have been over the life of the
settlement, but perhaps more likely, it dates from a short time at the end of the occupation
period. Burial within a currently occupied settlement was not usual Maori practice.

It could be argued that the pit burials date from a short period immediately before or
even during the Nga Puhi attack of late 1821. The general insecurity of the time, perhaps
even a Seige of the settlement, may have resulted in the unusual step being taken of burying
dead within the pa. This also provides an explanation for the subsequent exhumations. It
is possible the burials were uplifted for re-interment elsewhere, but the evidence points to
a less respectful purpose. In burial pits 6, 7, 8 and 9 there were a few bones left at the bottom
in a soil heavily impregnated with red ochre. The fact that not all bones were uplifted

Table 6. Basic data on burials uncovered at the 1987 and 1988 Raupa excavations.

Burial Sex Age — Burial form Ochre Orientation Head facing

] - 11-13 Trussed - on Plentiful West North
left side

2 M? adult Secondary Some East North

3(A) F 30-40 Knees drawn Abundant West South
up lying
under 3B

(B) M? 25-30 Onright side Abundant West South

- knees drawn
up. Over 3A

- M 18-19 Trussed-lying Absent SW -
on front. Head
missing

5 F 16-18 Extended Absent NE North

90 PRICKETT

strongly suggests that this was not a matter of re-interment elsewhere, but that the remains
were plundered out of disrespect. The use to which they were then put is suggested by the
shattered long bones in pit 7 which are almost certainly industrial waste. The use of human
bone for manufacturing purposes was not infrequently a final act in the defeat of an enemy
(see Vayda 1960:94-97).

A second group of burials dates from after abandonment of the site. All were dug
through the last occupation surface and so post-date the Raupa settlement. Also separating
them as a group from Burial 3 and the other probable burial pits was the lack of ochre
associated with the bones, only Burial | being similar to the earlier interments in this
respect. The fact that the ten burials or presumed burials of both periods are well separated
throughout the excavated area suggests that the location of the earlier group may have been
known to those who buried the latter. Thus, only a few years - or evena few weeks or months
- may have elapsed during which all the burials were put in place.

Burials 1, 2, 4 and 5 were in shallow scoops or pits, or in the last case was probably
laid on the ground to be covered by a few centimetres of soil. The range of practices
illustrated by these four finds may not indicate changes over time so much as differing
preferences by whanau decision makers or a variety of practical demands. As an example
of the latter, it is possible the secondary burial (2) was of an individual who died elsewhere,
to be brought back to Raupa for final interment.

Itis a surprise that there is no uniformity apparent in the burial style. Male and female,
young and old are laid to rest in as great a number of ways as there are individuals. Only
the distinction between the pit double burial and other presumed pit burials of the Raupa
occupation period and the variety of shallow burials which came after offers a pattern. Later
burials mostly lack ochre whereas the presumed pit burials of what may have been only a
few years before are marked by great quantities of it throughout the lower pit fill and in the
case of Burial 3 on the bones themselves.

The variety of burial practice and decreasing use of ochre to paint bodies may reflect
the uncertainty of Maori society in the first half of the 19th century. Musket wars between
the tribes and the incursion of Pakeha culture in all its commercial, religious and
technological guises may have upset old and agreed ways of doing things. Alternatively
there may always have been flexibility in these matters: the variety of burial arrangements
at the Washpool site in the Wairarapa (Leach and Leach 1979) which date from the 12th
or 13th century suggest there was nothing new in this.

Adzes

In the two seasons work at Raupa a total of eight adzes or substantial parts thereof were
found, along with six small polished fragments. Thus as many as fourteen adzes may be
represented altogether. All are, or appear to be, of the common rectangular, rounded
rectangular or sub-triangular Duff (1956: 166) Type 2B form, the most abundant adze of late
sites, especially in the North Island, Two more broken adzes of similar form were found
in 1984 by.Phillips (1986:102, 105).

RAUPA 9]

Probable
Eo early burial

1 (1987) BH

t 2 A
Early group. burials ep
KR OOD
REE

" Late group’ burials

Fig. 81. Plan showing distribution of ‘early’ and ‘late’ burials at Raupa.

The predominance of Waiheke Group greywacke (Schofield 1967) as adze raw
material was remarked upon in the report on the 1987 work (Prickett 1990:145, 147). All
eight further pieces are of the same stone, only one adze of the 1987 season being of a
different raw material, in this case basalt, probably from the Tahanga source on Coromandel
Peninsula (Prickett 1990:Fig. 38). The same greywacke is abundant among adzes from the
Oruarangi site where 14 of 17 examples listed by Fisher (1936:18) are identified as such.
The raw material of the other three adzes listed by Fisher is not given. The Raupa adzes
and fragments show a range of greywackes including very fine grained green material,
relatively coarse grey, and the stone, sometimes called ‘chipwacke’ (see Schofield 1967),
which is in this case a fine-grained green greywacke with small black inclusions.

Waiheke Group greywacke is shown by Schofield (1967) to occur in the inner Hauraki
Gulf islands of Rakino, Motutapu, Waiheke and Ponui, throughout the Hunua Ranges and
in the Hapuakohe Range of hills west of the Hauraki Plains. A more detailed recent map

9? PRICKETT

gives the distribution of greywacke and related rocks on Motutapu and Rakino Islands, the
east end of Waiheke and the Whitford district (Kermode 1991), Field observation suggests
the high quality green material probably came from the inner Hauraki Gulf islands.

Grey Waiheke Group stone is common in the river beds along the western shore of the
Firth of Thames. Schofield (1967) makes an interesting comment in noting that the Manaia
Hill Group of rocks, which are widespread on the Coromandel Peninsula and also west of
the Hauraki Plains, are ““...characterised by “chipwackes” (sandstones with angular chips
of argillite)”. These rocks may thus be the source of the chipwacke adzes in the Raupa
collection.

Among the Raupa artefacts the patu onewa found in three pieces in the 1987 and 1988
seasons is also made of greywacke, again of the chipwacke variety. Greywacke was also
present in the form of large water-rolled anvil stones and was present among oven stones,

Patu muka

A total of four handle and two beating parts of six separate patu muka were found in
Areas IV, VI, VI and X. The locations are at the rear of the Area III house and east of the
house where the pieces may be associated generally with the varied activity debris
concentrated here. A glance at Figs 24-27 in this account and Figs 41-42 in the 1987 report
(Prickett 1990:146) shows a variety of handle forms, and beating ends which conform
closely to the different shapes presented by the utilised water worn stones. The preferred
stone (five of six items) is grey andesite. Phillips (1986:102) reports another patu muka
found in her ‘Area 3’ (Trench E),

Patu muka were used in pounding stripped flax leaves to produce a soft clean fibre for
prestige woven cloaks. The thick fibre hanks were laid on a smooth boulder, such as were
found in Area VII especially, and beaten until soft enough to use. Patu muka breakage was
considerable if we are to judge from the very many fragments found at Oruarangi. Six
broken pieces at Raupa, and no complete item, support the conclusion from the larger
Oruarangi collection.

Nephrite

Seven of a total of nine nephrite items or fragments found in the 1987 and 1988
excavations are described in the section which covers Areas VII, [X-XI in this report. The
concentration of these items to the east of the Area III house mirrors the distribution of
obsidian, chert and other material. Clearly it reflects the use of this place as an important
activity area.

The three pendants or parts thereof represent a range of types including ear pendant
of kuru or kapeu form, a free-form, or adze-shaped pendant and a well shaped triangular
item which is related to the ancestral Polynesian whale-tooth ornament and its many
derivatives. Polished fragments and one complete item represent as many as five small
chisels used for fine wood working.

Bone artefacts

RAUPA 93

Bone artefacts found in the two years excavation include pieces of five tattoo chisels,
two bird spears, two fishhook fragments, two needle pieces, and five pieces of sawn or
otherwise worked bone. Phillips (1986:102, 105) found a bone awl, toggle and drilled dog
tooth in her ‘Area 3’, Trench E. Four of the five tattooing chisels were found together in
Area XI, metre square AS (Figs 39-42), showing that these items were not accidently lost
but carefully placed for later recovery. Similarly, four of the pieces of worked bone (Figs
38, 43-45) were found in Area IX at the south-east corner of the excavation; the other was
from the 1987 Area II (Prickett 1990:108). Just why the worked bone was together in Area
IX is not clear, it may be entirely coincidence.

Bone artefact finds represent a range of activities: tattooing, fishing, birding and
sewing. Mammal and bird bone is used. As with other artefactual material it is surprising
how little was found; the artefactual productivity of Oruarangi appears yet more remarkable.

Hoanga

Grindstones were concentrated in Area VI behind the house, and in Areas IV, VII and
XI, east of the building. The majority of pieces were derived from flaggy andesite such as
are illustrated in Figs 48-51 and most were used in kokowai grinding, Kokowai was also
prepared on water-rolled stones (Figs 54-58) and in one case on the surface of a broken patu
muka handle (see Fig. 24), Fragments of kokowai might also be used directly as in polishing
and colouring items: examples are described in the Areas VII, [X-XI section in this report.

Other hoanga were used in grinding and polishing stone tools, the most interesting
example bearing an incised design on one surface (Prickett 1990:147). Unlike the natural
slabs used in kokowai preparation this is made of fine-grained sandstone. Two more
examples of much used sandstone hoanga were found in Area VII and are pictured here in
Figs 52 and 53, In Area III at the rear wall of the house was found a piece of pumice much
used for polishing on one flat surface (Prickett 1990:123).

Material of European origin

Half a century before Marsden visited Raupa in 1820, Cook penetrated the Waihou
River, pushing upstream in a pinnace and long boat to within a few kilometres of the Waihou
and Ohinemuri confluence (Beaglehole 1955:206). In the decades following there was
considerable European contact with the district, mostly to obtain spars for which the tall
and straight kahikatea so admired by Cook was well suited. In 1801 a timber cutting gang
from the ‘Royal Admiral’ went so far as to erect a 2 m high defensive stockade on the bank
of the Waihou River (Smith 1813). To the north, in the Bay of Islands, there was a growing
European population by 1820. Thus Maori people living along the Waihou River had
considerable opportunity for contact with Pakeha many years before Raupa developed into
the substantial settlement described by Marsden. This is notreflected in the few items dating
from the Maori settlement which are of European origin.

Most material of European origin described in reports of the two seasons work clearly
dates from after Maori occupation, being recovered from surface silt or upper mixed layers.

94 PRICKETT

Only a few items can be argued to belong to the period of Maori settlement at Raupa,
probably in the early 19th century. These are a glass wine bottle fragment in Area IV
(Prickett 1990:134-135), a piece of rusted iron from within a posthole in Area VII anda clay
pipe bowl fragment from Area X. Pig bone was identified in Area II and possibly Area I
(Prickett 1990:88, 107). A blue glass bead found in disturbed ground by Phillips (1986:99)
may also relate to the Maori occupation.

It was a surprise that so little material was found to reflect many decades of contact
in a part of northern New Zealand long frequented by European visitors. Archaeological
evidence at Raupa suggests that stone adzes were still used for wood-working, obsidian and
chert knives for cutting and scraping tasks, and bone fishhooks, birdspears and needles still
occupied their place in everday life. It could be argued that valuable iron and steel tools were
more carefully looked after and thus were not left at the site; it is the experience of
archaeology, however, that even the most valuable items might sometimes be lost. As late
as 1820 European material culture apparently had made little impact on Ngati Tamatera
and other people living along the Waihou River.

Obsidian

In two seasons work a total of 3588 pieces of obsidian weighing 19558.1 g were
recovered at the Raupa site. This very large sample of almost 20 kg deserves a great deal
more study than is possible in the present report, nonetheless some general aspects of the
collection may be described and conclusions drawn.

It is interesting to speculate on the total volume of obsidian in the Raupa site, with ca.
20 kg coming from just two percent of the area. What is more, from the evidence of artefacts
recovered, Raupa had nothing like the intensity or longevity of occupation as Oruarangi and
possibly other Hauraki sites as well. Thus the total volume of obsidian transported to this
relatively small group Waihou River sites may have been very considerable indeed.

Almost all Raupa obsidian came from Mayor Island, 95.5% by weight coming from
that source. Of the remainder, 3.3% came from the Waimata Stream, Waihi, source which
has been described by Moore and Coster (1989) and 1.2% was pale grey in colour and
sourced only to a variety of Coromandel, Great Barrier Island or central North Island
localities. One small piece originated in Whangamata, Coromandel, and three fragments
of black obsidian in Area IV came from a source unknown.

The preference for Mayor Island obsidian is very marked, the more so when one
considers that the alternative Waihi source is only 20 km from Raupa and is close by the
shortest route to Mayor Island. The streambed origin of Waihi obsidian is reflected in the
water-rolled, sometimes pitted, surfaces that are common among that material. This
demonstrates the use of small cobbles which were not as easily worked or as useful as the
much larger blocks available from Mayor Island.

The distribution of obsidian in excavated parts of Raupa shows an interesting pattern.
In Areas IT and VIII there was virtually none, and the little that was found in Area V probably
reflects its location at the edge of the intensive working a few metres to the south. Also, Area

RAUPA 95

I had comparatively little obsidian which may be scattered from a more intensive working
area nearby, or could simply reflect very occasional working or use of the stone in this area.
It is the large area excavated about the 1987 house that deserves the most attention for its
distribution of obsidian.

Fig. 82 summarises obsidian distribution in Areas III, [V, VI, VII, LX, X and XI. For
the purpose of this analysis Levels I and II material is combined, at least partly because
of the difficulty in accurately assigning material in some parts of the excavation. Most stone
came from Level II, but much of that assigned to Level I may actually belong with the later
occupation (as in Area IV, see Prickett 1990:130). Also, Area III was not fully excavated
so more material here may increase the present figures. Nonetheless, the distribution as
shown probably reflects fairly accurately the actual scatter of obsidian in the upper
occupation level.

Within the house itself obsidian was scattered about the floor, except that less was
found in front of the rear wall than elsewhere (see Prickett 1990:119). This compares with
the situation in the Moikau house, Palliser Bay, where obsidian, in a much smaller sample,
was more plentiful at the rear than elsewhere in the building (Prickett 1979:42). At Raupa,
however, we cannot be certain how this distribution reflects use inside the house since it
may have been altered by activity after the building was destroyed by fire. At the rear of
the house (Area VI) there was little obsidian use.

The major concentration of obsidian was east of the front part of the house. This was
obviously an important area for preparing obsidian tools and probably also for using the
sharp-edged stone in tasks such as cutting flax and other fibre and working wood. Obsidian
was also used for processing food, such as cutting up birds, fish and other animals, but it
seems unlikely that this area, away from cooking ovens and close to the entry of an
important house, would have been used for this purpose. Fig, 82 shows clearly that the
excavated area covers the focus of whatever activity was taking place, with obsidian
densities decreasing to the outer margins of the excavation. At the south-east corner in Area
IX what little obsidian was found probably marks the outer edge of this focus of intensive
use.

Analysis of the 1988 season obsidian shows that approximately 60% of all pieces are
flakes struck on purpose or accidentally from larger blocks. Most of the remainder are
shatter pieces, with ca. 1.5% cores or core remnants. Approximately 30% of flake and
shatter pieces alike have been used for cutting or scraping tasks which have left evidence
in the form of edge-damaged, battered or notched margins. The very large obsidian
assemblage from Raupa warrants more detailed analysis.

Chert

A total of 1258 pieces of chert weighing 1 6354.0 g were recovered in two seasons work
at Raupa. By weight more than half of this came from Areas VII, [X-XI and almost all of
the remainder from Areas III, [V and VI. The outlying Areas I, II, V and VII produced only
304.9 g between them, 82% of it from Area I.

Most of the chert was ofa recognisably similar material, predominantly red, white and
brown in colour, but which could include also a variety of other colours such as grey and

96 PRICKETT

Fig. 82. Plan showing obsidian distribution in Areas III, IV, VI, VII, LX, X and XI.
Numbers give weight of material in metre squares - ‘1’ = 100-200 g, ete. Diagonal line
marks metre squares of 50-100 g.
yellow. This chert could vary greatly in quality from fine-grained material to coarse
unusable stone. Some had water-rolled surfaces, suggesting that it came from stream beds
as well as primary deposits, almost certainly in the nearby Coromandel Range. There was

in addition a wide variety of other cherts and a few pieces of chalcédony and jasper.

In his summary of New Zealand chert sources Moore (1977:55) shows the Coromandel
Peninsula and ranges south to the Ohinemuri River to be made up almost entirely of chert
bearing formations. On the western side of the Hauraki Plains and on Hauraki Gulf islands
chert is associated with the Waiheke Group greywacke from which most Raupa adzes have
been made. Thus the Raupa material may also have come from there. With such abundant
sources nearby it seems unlikely that it came from further afield.

Chert was probably used for much the same cutting and scraping jobs as was obsidian,
although it is a tougher, less brittle, material which might be an advantage in some tasks.
It is also, however, coarser and often less easy to flake than obsidian. Much of the recorded
weight of chert comprises large blocks of variable or poor quality material which have in
all probability been discarded as not worth further attempts to strike off useful sharp-edged
tools.

RAUPA 97

An archaeological history of Raupa

The report on the 1987 excavations concluded with an outline history of Raupa based
on the results of that year’s work (Prickett 1990:148-150). The second season’s excavations
confirm some of the earlier conclusions and add to our knowledge of the human history of
the site.

In the 1987 discussion it is suggested there was evidence in Areas I and V for reversion
of these parts of the site to open ground from previous use for structures or cooking activity.
In Area II a storage pit, and many postholes to show structures, are buried by fill in which
are shallow cooking ovens. The 1988 work showed a similar trend in Area VI with early
cooking ovens covered by fill on which there was a featureless occupation surface. At the
south-east of the combined Areas VII, IX-XI excavation the Level I structure, probably a
small dwelling, was covered in crushed midden, black soil and oven stones on which surface
there were a few postholes.

The conclusion to be drawn from all this is that the last occupation at Raupa in all
likelihood involved a smaller population than the earlier settlement. Whether this was at
the time of the 1820 visit by Marsden, or later, is unclear. Marsden’s account suggests a large
and thriving settlement, but it is possible that some stage before his visit there was an even
larger settlement here, utilising more fully the available space within the bend of the river
and behind the defensive lines across the isthmus. Alternatively the archaeology may
document a decline after Marsden’s visit.

The radiocarbon dates suggest that the earlier occupation deposits at Raupa were at
most 100 years older than the 1820 settlement. The age difference may have been much
less. Nowhere in the site was there a suggestion of soil formation between earlier and later
occupation deposits. Silt layers are likely to relate to particular flood events and may mark
the passage of no more than a day or two. The traditional reference to Raupa being visited
as early as “A.D. 1600” (see Kelly 1945:210) may refer to a settlement at the excavated site,
but no sure indication of occupation at this early date was found.

In the historic outline preceding the 1987 report (Prickett 1990:73 and 75) the fate of
Raupa following Marsden’s visit is briefly outlined. The Nga Puhi who took Te Totara in
late 1821 went on to Raupa where, according to a 1893 Land Court witness, they were
defeated (Rihitoto 1893).

It was suggested in the conclusion to the first season’s work that the Nga Puhi attack
on Raupa may have had a very different outcome. The large house burnt to the ground, its
floor partly dug over for cooking ovens, with fragmentary human bone trod into the old
ground surface, together indicate a violent end to that phase of Raupa settlement. Further
evidence is now available. In Areas VII and XI, pieces of human bone were found scattered
about the occupation surface, a wooden bowl and cloaks were burned in shallow scoops,
and widely scattered fragments of patu onewa were found to suggest that the weapon may
have been broken in fighting rather than by fire as was suggested in the 1987 report. Also,
as has already been argued, it is possible that some of the burials were uplifted out of
disrespect and ritual defilement.

98 PRICKETT

Whatever the actual event, an argument can be made on the archaeological evidence
alone that there was a violent end to the later phase of occupation at Raupa which involved
at least some deliberate destruction, defilement and also cannibalism. That these should
mark an unknown event some time after the Nga Puhi attack of late 1821 seems unlikely.

The Raupa settlement above all illustrates something of the political uncertainly of the
early decades of the 19th century. The large and thriving settlement seen by Marsden may
have included people who had moved up the Waihou River to escape Nga Puhi and other
raids from the north. It was primarily, however, a well established pa of the Ngati Tamatera
people in which Marsden noted houses “...much larger and better built...” than any he had
seen previously in New Zealand (Elder 1932:255). These large houses may reflect a society
not yet seriously disrupted by warfare which was to engulf New Zealand in the following
two decades. At Raupa the disruption - and indeed destruction - is clear in the archaeological
record.

After the Nga Puhi raid of 1821 the Hauraki tribes retired southwards for fear ofa repeat
visit. For some years they lived among Ngati Haua and Waikato tribes until in 1830 the
battle of Taumatawiwi was fought between them and the host tribes. This fiercely fought
encounter was not decisive but the Hauraki people agreed afterwards to return to their
homes. Just how Raupa fits into these events is not entirely clear. It is likely that the pa was
substantially if not completely abandoned soon after the Nga Puhi raid of late 1821. Before
the Hauraki tribes accepted Christianity - and consequently very different burial practices
to those in evidence at the site - Raupa was an abandoned settlement deemed suitable as
a burial ground for its associations with the recent past.

One other general point needs to be made. In the above discussion regarding items of
European origin it is remarked that European material culture had little apparent impact on
the technology and economy of the people at Raupa as late as 1820. This result was counter
to expectations, based on the supposed devastating impact of new technologies and material
culture on the old Maori world. At Raupa there was little archaeological evidence of this
impact. The conclusion must be that whatever imported items were present, the people of
Raupa still depended largely upon traditional ways of doing things. While it may be
suggested that the new material culture was too valued to be discarded or lost, nonetheless
obsidian and chert knife technology, stone adzes and other stone and bone tools show that
these anyway were still important.

Acknowledgements. Thanks are due to the Ngati Tamatera of Paeroa and surrounding districts for
accommodation at Te Pai o Hauraki marae and for support in this project which I hope has added
something to knowledge and understanding of their past. In particular I would like to acknowledge
the practical assistance and unfailing kindness of the late Huhurere Tukukino, Dick Rakena, Mate
Royal and Winnie Hutchinson.

Major funding came from the New Zealand Historic Places Trust with assistance also from the
Labour Department. The Anthropology Department, University of Auckland, provided excavation
equipment and laboratory space. The Auckland Institute and Museum gave essential support for the
two year project.

Thanks are also due to the large number of people who gave their time and effort to take part
in the ‘dig’. I hope I will be excused if I do not list everyone here. I must, however, thank in particular

RAUPA 99

area supervisors Anne Leahy, Jolanda Cupido, Brian Wedlock, Tony Walton, Nancy Tayles and
Warren Gumbley. Pat Stodart provided essential help in organising gear and making sure the
excavation ran smoothly. Careful work with the burials and the recovery of some information before
the pits were refilled was the responsibility of Nancy Tayles and Yuriko Igarashi. Nancy Tayles also
prepared a report on the burials which is the basis of the description in this report.

Analysis of stone material was carried out largely by Kath Prickett with specialist help from
Dante Bonica and Phil Moore. Liz Hudson examined the fragmentary and crushed human and dog
bone from Areas VII and XI. Rod Wallace of the University of Auckland Anthropology Department
identified charcoal in the radiocarbon samples. Alan Hogg and Tom Higham of the University of
Waikato Radiocarbon Dating Laboratory provided the radiocarbon dates and helped in their
analysis. Dilys Johns of the University of Auckland Anthropology Department came to the site at
a moment’s notice to uplift the Area VII wood and fibre deposit, and later conserved the material.
Maureen Lander made an excellent photographic record of the fibrework, some of the results of
which are included in her report which is published in this volume, Joan Lawrence drew the
illustrations. Mary Best typed the manuscript.

REFERENCES

ANON
1884 Plan of roads through blocks Rawhitiroa, Parahumati ... Map no. 3421, Dept of Survey
and Land Information, Hamilton.

BEAGLEHOLE, J.C.
1955 The Voyage of the Endeavour 1768-1771. Cambridge, Cambridge University Press for
the Hakluyt Society.

Durr, R.S.
1956 The Moa-Hunter Period of Maori Culture. Wellington, Government Printer.

Exper, J.R,
1932 The Letters and Journals of Samuel Marsden 1765-1838. Dunedin, Coulls Somerville
Wilkie and A.H. Reed.

FIsHER, V.F.
1934 The material culture of Oruarangi, Matatoki, Thames.
1, Bone ornaments and implements. Rec. Auckland Inst. Mus. 1:275-286.

1936 The material culture of Oruarangi, Matatoki, Thames.
3. Stone implements and ornaments. Rec. Auckland Inst. Mus. 2:15-27.

HoucutTon, P.
1980 The First New Zealanders. Auckland, Hodder and Stoughton.

HouaGutTon, P., B.F. LEAcH AnD D.G. SUTTON
1975 The estimation of stature of prehistoric Polynesians in New Zealand. J. Polynes. Soc.
84:325-336,

KELLY, L.G,
1945 Tapuariki and Raupa. J. Polynes. Soc. 54:199-211.
1949 = Tainui. Polynes. Soc. Memoir 25.

Kermope, L.O.
1991 Whangaparaoa-Auckland, Infomap 290, Sheet R10/11 1:100000. Wellington, Dept of

100 pRICKETT

Survey and Land Information.

Law, R.G.
1980 Regional variations in Maori greenstone pendants of the kuru style. Rec. A uckland Inst.
Mus. 16: 63-75.
Leacu, B.F. and H.M. Leacu
1979 Burial positions and orientations in Palliser Bay. In, Prehistoric Man in Palliser Bay,
B.F. and H.M. Leach (Eds). Bull. Nat. Mus. N.Z. 21:205-213.

McFapcen, B. and M.R. MANNING
1990 Calibrating New Zealand radiocarbon dates of marine shells. Radiocarbon 32:229-232.

Moore, P.R.
1977. The definition, distribution and sourcing of chert in New Zealand. N.Z. Archaeol. Assn
Newsl. 20:51-85.

Moore, P.R. and J. Coster
1989 Waihi obsidian. Archaeol. in N.Z. 32:26-30.

PHILLIps, C.
1986 Excavations at Raupa pa (N53/37) and Waiwhau village (N53/198), Paeroa, New
Zealand, in 1984. N.Z.J. Archaeol. 8:89-113.
1988 University of Auckland Field School excavations at Waiwhau, 1987. N.Z.J/. Archaeol.
10:53-72.

Puituips, C. and R.C. GREEN
1991 Further archaeological investigations at the settlement of Waiwhau, Hauraki Plains.
Rec. Auckland Inst. Mus. 28:147-183.

PrickeTT, N.J.
1979 Prehistoric occupation in the Moikau Valley, Palliser Bay. In, Prehistoric Man in Palliser
Bay, B.F. and H.M. Leach (Eds), Bull. Nat. Mus. N.Z. 21:29-47.
1982 An archaeologists’ guide to the Maori dwelling. N.Z./. Archaeol. 4:111-147.
1990 Archaeological excavations at Raupa: the 1987 season. Rec. Auckland Inst. Mus. 27:73-
153.

RIHITOTO
1893. Hauraki Minute Book Vol.30; Waikato-Maniapoto Land Court (Paeroa 12 April 1893).

SCHOFIELD, J.C.
1967. Sheet 3 Auckland (Ist ed.). Geological Map of New Zealand 1:250000. Wellington,
Dept. Sci. Ind. Res.

SKINNER, H.D.
1974. Maori amulets. Pp. 45-98 in, Comparatively Speaking, P. Gathercole, B.F. and H.M.
Leach (Eds). Dunedin, University of Otago Press.

SmitH, W.
1813. Journal of a Voyage in the Missionary Ship Duff. New York, Collins.

Sturver, M., G.W. PEARSON and T. BRAZIUNAS
1986 Radiocarbon age calibration of marine samples back to 9000 cal. yr BP. Radiocarbon

RAUPA 101

28:980-1021.

TE RANGi HirRoA
1949 The Coming of the Maori. Wellington, Maori Purposes Fund Board.

Vaypa, A.P.
1960 Maori Warfare. Wellington, Reed for the Polynes. Soc.

Warp, G.K. and S.R. WILSON
1978 Procedures for comparing and combining radiocarbon age determinations: a critique.

Archaeometry 20:19-31.

NIGEL PRICKETT, Auckland Institute and Museum, Private Bag 92018, Auckland.

A NEW LOWER MIOCENE SPECIES OF ANODONTIA
(MOLLUSCA: BIVALVIA)

MICHAEL K. EAGLE

AUCKLAND INSTITUTE AND MUSEUM

Abstract. A new species, Anodontia (Anodontia) waharoaensis, is described from the
Lower Miocene Waitakere Group, Motutara, West Auckland. It is the first record of the
genus from New Zealand.

Numerous trips to the coastline of Motutara, between Otakimiro Point,
Muriwai, and Raetahinga Point, Motutara, have yielded many varied and unusual
macrofossils (Powell 1935, Hayward 1976a). A partial crystalline shell and incomplete
internal cast of a large bivalve were collected by the author from Tirikohua Point on
22 June 1990. An expedition further south to Te Waharoa Bay on 23 and 24 May
1992, located several bedding planes with internal casts and a complete, crystallised,
articulated specimen of the same mollusc in situ. Site locality numbers are those of
the Geological Society of New Zealand’s archival Fossil Record File.

GEOLOGICAL SETTING (Fig. 1)

The two fossil sites are located within the Waitakere Group, Manukau Subgroup,
Tirikohua Formation (Hayward 1976b). They are: 1. Tirikohua Point Site, Q11/f45,
Bartrum Member; and 2. Te Waharoa Bay Site, Q11/f52, Otakimiro Member, Both
Bartrum and Otakimiro Members were deposited in a submarine canyon and consist of
thick, massive, fine pebble-bearing volcanic grit or coarse sandstone containing lenses of
cross-bedded volcaniclastic sandstone (Hayward 1976a). The stratigraphy of the Otakimiro
Member, from which the holotype came, is shown in Fig. 1. For stratigraphy of the Bartrum
Member see Hayward (1983).

SYSTEMATICS
Class Bivalvia
Subclass HETERODONTA Neumayr, 1884
Order VENEROIDEA H. Adams & A. Adams, 1856
Superfamily LUCINACEA Fleming, 1828
Family LUCINIDAE Fleming, 1828
Subfamily MILTHINAE Chavan, 1969
Genus Anodontia Link, 1807
Subgenus Anodontia Link, 1807
Type species : A. alba (= Venus edentula, Linne, 1758; OD).

Recent, Indian Ocean.

Rec. Auckland Inst. Mus. 29: 103-111 1992

104 EAGLE

100m TIRIKOHUA €:):
” POINT ‘NO

O*
KIRIKIRI BEACH
mn

\
Gauss TE WAHAROAS3
@rs2 POINT «

a pine

STRATIGRAPHIC
SECTION
FOSSIL LOCALITY

KAIHU GROUP

_|[-_—SMitiwai Sand qm
WAITAKERE GROUP

MANUKAU SUBGROUP
NIHOTUPU FORMATION mn

eects sandstone

oan volcanic breccia
22g slumped horizon
Wy cross-bedded

ee fayit — NOrizon |FE==E==|TIRIKOHUA FORMATION mt
G fossil locality Otakimiro Member mto
0 0.5m Te Waharoa Conglomerate Member mtt

__ Bartrum Member mtb
PIHA FORMATION mp

Fig. 1. Geological map of the northern part of Te Waharoa Bay within the area known
as Motutara showing the fossil type locality and stratigraphic column for Te
Waharoa Point (after Hayward 1983).

Anodontia (Anodontia) waharoaensis n. sp. (Figs. 1-6, 9, 10).
MATERIAL

Holotype. AK72308 (Auckland Institute and Museum), complete articulated crystalline
bivalve, Q11/f52.

Paratypes. AK72309 (Auckland Institute and Museum), 6 internal casts or impressions,
Q11/f52. ME9002 (author’s collection) 1 internal cast, Q11/f52. L3862 (University of
Auckland), 2 internal casts or impressions, Q11/f45.

TYPE LOCALITY

Fossil record file number Q11/f52; grid reference Q11/382816 (NZMS 260
1:50 000 map), Te Waharoa Bay, Motutara, West Coast, Auckland,

ANODONTIA 105

DESCRIPTION

Shell very large, subcircular, slightly longer than high, strongly inflated and
globose, subspherical, slightly truncate posteriorly; equivalve; beak slightly anterior of
midline, thick and solid. Beak prosogyrous, low and inconspicuous. Shell surface has
weak, irregular, comarginal grooves and low growth lamellae. A weakly defined shallow
sulcus and low ridge continuing forward from below the beak define a narrow, slightly
raised postero-dorsal area on which the comarginal lamellae are less consistently developed
than elsewhere. Lunule tiny, deep.

Ligamental groove oblique, sunken, lanceolate, with inward sloping walls. The
hingeline is gently curved; postero-ventral margin broadly rounded. Shell interior and
internal casts prove lower, inner section of the anterior adductor muscle scar to be long,
lying well above the pallial line and elongated vertically. Posterior adductor
muscle scar trapezoidal, broader, situated at a similar distance above pallial line than
anterior scar. Internal casts possess an entire pallial line; whole shell interior dimpled

Figs. 2-5. The holotype AK72308. 2. Right valve showing slightly truncate posterior

shape and large size. 3. Left valve showing weak, irregular, comarginal grooves and

low growth lamellae. 4. Conjoined valves of holotype in dorsal view showing

prosogyrous beaks, weakly defined shallow sulcus and low ridge continuing forward

from below the beaks. 5, Reconstructed internal right valve showing both anterior and

posterior adductor muscle scars, a dimpled and rugose shell interior, and an entire
pallial line.

106 EAGLE

Fig. 6. Partial left valve steinkern of the paratype L3862b showing the long anterior

adductor muscle scar elongated vertically well above the pallial line (arrowed), and the

zone between the pallial line and the ventral margin covered in fine, irregular, radial
lirae c.1 mm apart.

0) 20mm
(| Seeiae

20mm

Figs. 7-9. Dorsal views of one Recent and two fossil species of

Anodontia (Anodontia) comparing conjoined valves, beak configuration, and inflation.

7. A. (A.) inflata (Eocene). 8. A. (A.) edentuloides (Recent). 9. A. (A.) waharoaensis
(Miocene).

ANODONTIA 107

and rugose; ventral margin entire; zone between pallial line and margin bearing fine
irregular radial lirae 1 mm apart, stronger and more regular near margin.

Measurements of the holotype (AK72308) are: length 123 mm; height 123 mm;
convexity 48 mm; postumbonal length 65 mm; anterior-dorsal apical inclination 25°.
Because the paratypes are incomplete they cannot be measured accurately.

AGE
Middle-Altonian (Pl) (Burdigalian), early Miocene (Hayward 1983).
ETYMOLOGY
Named after the type locality, Te Waharoa Bay.
FAUNAL ASSOCIATIONS

Powell (1935) and Hayward (1976c) described a small but highly distinctive
bathyal molluscan fauna from rocks north of the present fossil localities. Several
taxa discovered there occur in situ with A. (A4.) waharoaensis (Table 1). Saccella is an
accepted outer shelf to upper bathyal genus, as is the gastropod Bathytoma, and their
association with the genus Amygdalum, of which there are no recent records from the
seas around New Zealand, confirms this paleodepth. Amygdalum now lives in tropical
and warm temperate waters, byssally attached to a hard substrate mostly in outer
shelf to bathyal depths (Beu et al. 1990), suggesting a similar environment in the mid-
Altonian at Te Waharoa.

Table 1. Systematic list of fossil fauna found in association with 4. (A.) waharoaensis.
Nomenclature follows Beu et al. (1990),

SS EE

MOLLUSCA

BIVALVIA
NUCULANIDAE Saccella motutarensis (Powell, 1935)
MODIOLINAE Amygdalum dolichum (Suter, 1917)
LIMIDAE Lima colorata Hutton, 1873
LUCINIDAE Lucinoma taylori (Powell, 1935)
TELLINIDAE Elliptotellina protensa (Powell, 1935)

GASTROPODA

ACMAEIDAE Pectinodonta conformis (Marwick, 1931)
TURRITELLIDAE Zeacolpus sp.
TURRIDAE Bathytoma michelsoni Powell, 1935
CUVIERINIDAE Vaginella inflata Hayward, 1981

ANNELIDA

POLYCHAETA Protula sp.

108 EAGLE

Figs. 10-13. Comparison of one Recent and three fossil species of Anodontia

(Anodontia) showing differences in valve shape, anterio-dorsal angle of apical

inclination, and comarginal lamellae. 10. Holotype of A. (4.) waharoaensis (fossil). 11.

Holotype of A. (A.) inflata (Kern County, Southern California - fossil). 12. Holotype

of A. (A.) sphericula (Edithburg, South Australia - fossil). 13. A. (A.) edentuloides
(Magdalena Bay, California - Recent).

DISCUSSION

This is the first record of the genus Anodontia (Anodontia) from New Zealand.
Fleming (1950) described a new lucinid, Loripinus arapaoa, from the Pahi Greensands,
Pahi Peninusula, North Auckland. This Eocene (Bortonian) mollusc was tentively allocated
by Fleming to the subgenus Eophysema Stewart, because of its shape and fine radial
sculpturing. Following Chavan’s (1969) revision of the Lucinidae, itis clear that Fleming’s
species should now be referred to Pegophysema (Eophysema) arapaoa rather than
?Loripinus arapaoa (Beu et al, 1990).

Fleming (1950: 245), mentions that “additional species [of lucinid] as yet
undescribed, occur in the Miocene [of New Zealand] and reach over twice the size of
the recent species”. Anodontia (Anodontia) waharoaensis n.sp. 1s larger than both the
three known recent species and the two fossil circum-Pacific species. It possess the
greatest anterior dorsal apical inclination (25°) and is the most posterior- truncate in shape.

The Recent type species Anodontia (Anodontia) alba (Linne, 1792) is only half
the size of the new species (length 51 mm, height 44 mm, convexity 31 mm), is more
circular, and is not as inflated. It is similarly heavy and fairly strong, but with much
weaker, irregular, concentric growth lines. The postero-dorsal area is not as raised.
The anterior dorsal margin incorporating a reduced ‘wing’ is convex and not straight as in
A.(A.) waharoaensis. Internal adductor muscle scars also differ in size and shape; the

ANODONTIA 109

posterior scar is proportionally smaller and more oval than the anterior scar which is longer
and not as wide.

Anodontia (Anodontia) stearnsiana (Oyama et al., 1960), differs from the new
species by being only half the shell size (length 62 mm, height 56 mm, convexity 39
mm). It has a more prominent beak apex, subalate anterior and posterior dorsal margins,
and is more circular in shape. The posterior area is not as raised in A. (A.) waharoaensis,
and the irregular comarginal growth lines are finer.

Anodontia (Anodontia) edentuloides (Verrill, 1870) is subglobose and convex,,
with a comparably thin shell (Figs. 8 & 13). It is approximately one third the size of the
new species (length 45 mm, height 38 mm, convexity 27 mm). The irregular growth
lines differ in being ‘wrinkled’, however they are set off by deep marks indicating
growth resting stages, and are therefore ‘banded’. Minute radial striae may also be
seen. The pallial line differs from the new species in that it is covered with a punctated
calcareous coating. The ligamental groove is shorter and of all the specimens
compared, this species has the smallest anterior-dorsal apical inclination (17°).
The adductor muscle scars are not as long or broad as in the new species.

Darragh (1970) allocated Meretrix sphericula Basedow, 1902 (Miocene, South
Australia) to the genus Anodontia. Anodontia (Anodontia) sphericula (Fig. 12) 1s
different from A. (A.) waharoaensis in being more circular, having a surface
ornamentation of fine concentric threads between numerous, similar, coarse, and irregular
growth lines. On denuded shell areas, faint radial striae are more numerous
and prominent than those occasionally found on the new species. The lunule in A. (A.)
waharoaensis is tiny and deep, whereas in A. (A.) sphericula it is prominent and
lanceolate. The Australian fossil is approximately spherical (length 74 mm, height 67
mm, convexity 50 mm) and more regularly convex in appearance than the posteriorly
truncate A. (A.) waharoaensis. As it is comprised of two conjoined valves, the interior
of the A. (A) sphericula specimen is unknown.

Moore (1988) tentatively allocated the poorly preserved Eocene Phacoides (Callucina)
inflata (Fig. 7) from southern California to Anodontia (Anodontia) on the basis of the
rounded shape and very globose valves (length 66 mm, height 61 mm, convexity 43 mm).
Like A. (A.) waharoaensis, the remaining shell has no fine concentric ridges or lines, is
subcircular and broadly truncate with the beak slightly anterior to centre. The specimen
possesses a distinct, small, deeply impressed lunule,and a straight anterior dorsal
margin (see Figs. 7 & 8). Both species are ribbed internally by faint, irregular radial
ribbing, which crenulates the ventral edge in A. (A.) inflata but not in A, (A.)
waharoaensis. The new species is more truncate posteriorly, possesses a more
oblique, sunken, ligamental groove, is longer, higher, and more convex. The irregularly
spaced incremental growth lines are fewer and are coarser than those of A. (A.) inflata
(Figs. 10-13).

Anodontia (Anodontia) is found as a fossil in deep-water Eocene and younger
rocks in Europe, Asia, North America, the Pacific Ocean and Australia. Living taxa are
found in shallow, (15-100 m) mud-bottomed depths from North Carolina to Florida, in
the Gulf of Mexico, in the West Indies and Bermuda, in Magdalena Bay in South

110 EAGLE

California, from Kyushu to Honshu and in the Amami O shima Islands in Japan. It
appears therefore that this genus has changed its preferred bathymetric niche during
the Neogene.

Acknowledgements. | thank Brett Stephenson for his excellent photography and processing (Figs.
2-4, 6,9, and 10). Figs. 7 and 11 were reproduced from Moore (1988); Figs. 8 and 13 were reproduced
from Olson (1961) and Fig. 12 from Basedow (1902). I am grateful to Bruce Hayward, Anthony
Wright, Chris Hollis (Auckland Institute & Museum), and Jack Grant-Mackie (University of
Auckland), for assistance, helpful comments, and time taken in reading the draft manuscript.

REFERENCES

Basepow, H.
1902 Descriptions of new species of fossil Mollusca from the Miocene
Limestone near Edithburgh. Royal Society of South Australia 3: 130-131.

Bev, A.G., P.A. Maxwe i and R.C. Brazier
1990 Cenozoic Mollusca of New Zealand. New Zealand Geological Survey Paleontological
Bulletin 58: 1-518

CHEVAN, A.
1969 Lucinidae. InR.C. Moore & C. Teichert (eds.). Treatise of Invertebrate Paleontology.
Pt. N., Mollusca, 6, Bivalvia. The Geological Society of America and the University
of Kansas; N502-N504.

DarraGhi, T.A.
1970 Catalogue of Australian Tertiary Mollusca (except chitons). Memoirs of the National
Museum of Victoria 31: 125-212.

FLEMING, C.A,
1950 The molluscan fauna of the Pahi Greensands, North Auckland. Transactions of the
Royal Society of New Zealand 78: 236-250.

Haywarp, B.W.
1976a Lower Miocene geology and sedimentary history of the Muriwai - Te Waharoa
coastline, North Auckland, New Zealand. New Zealand Journal of Geology and
Geophysics 19: 639-662.

1976b ~—LLower Miocene stratigraphy and structure of the Waitakere Ranges, North Auckland,
New Zealand and the Waitakere Group (new). New Zealand Journal of Geology and
Geophysics 19: 871-895.

1976c Macropaleontology and paleoecology of the Waitakere Group (lower Miocene),
Waitakere Hills, Auckland. Tane 22: 177-206.

1983 Sheet Q11, Waitakere. Geological Map of New Zealand 1:50 000. New Zealand
Geological Survey, Department of Scientific and Industrial Research. Map (1 sheet)
and Notes. 28p.

ANODONTIA 111

Moorg, E.J.
1988 Tertiary marine pelecypods of California and Baja California: Lucinidae through
Chamidae. United States Geological Survey Professional Paper 1228-D: 1-46.

Oxson, A.A.
1961 Mollusks of the Tropical Eastern Pacific. Panamic - Pacific Pelecypoda.

Paleontological Research Institution, Ithaca. 574p.

PoweL_, A.W.B.
1935 Tertiary Mollusca from Motutara, West Coast, Auckland. Records of the Auckland
Institute and Museum 1: 327-340.

OyaMa, Katsura, MizuNA, ATSUYUKI, SAKAMOTO and ToRU
1960 Illustrated Handbook of Japanese Palaeogene Molluscs. Geological Survey of
Japan. 244p.

M.K. EAGLE, Auckland Institute and Museum, Private Bag 92018, Auckland.

PALEONTOLOGY AND PALEOECOLOGY OF EARLY
MIOCENE SEQUENCES IN HAYS AND TIPAKURI
STREAMS, NORTHERN HUNUA RANGES, AUCKLAND

MiIcHAEL EAGLE AND Bruce W. HAYWARD
AUCKLAND INSTITUTE AND MUSEUM

Abstract. One hundred and thirty-one macrofossil taxa (51 gastropods, 35 bivalves, 19
corals, 9 fish (otoliths), 4 scaphopods, 4 brachiopods, 4 barnacles, 3 polychaetes, |
shark (tooth), 1 echinoderm) are recorded from two rich, early Miocene (Otaian),
Waitemata Group fossil beds exposed in Hays and Tipakuri Streams in the northern
Hunua Ranges. Ninety-seven species of foraminifera (protozoan microfossils) are
recorded from six faunas recovered from the associated strata, These macro- and
microfossil faunas are used to interpret the paleoenvironmental history of the area.

The previously unrecorded Tipakuri Stream fauna in basal Waitemata Group
muddy sandstone (Kawau Subgroup) appears to be an in-situ assemblage thataccumulated
at mid to outer shelf depths during the early subsidence of the Waitemata Basin.

The Hays Stream macrofauna, from which 35 taxa had previously been recorded,
occurs in Waitemata Group strata transitional between the basal shallow water Kawau
Subgroup and the deep bathyal Warkworth Subgroup flysch that filled most of the
Waitemata Basin. The Hays Stream macrofauna contains a mixed assemblage that we
interpret to be derived from four distinct groups of communities: 1. intertidal and
shallow subtidal sand and mud communities; ii. intertidal and shallow subtidal rock and
coarse gravel communities; ili. inner to mid shelf silt, sand and fine gravel communities;
iv. outer shelf to upper bathyal mud and sand community. Shells, gravel, sand and mud
from these communities are inferred to have mixed together in a mass flow that carried
them down to deep upper bathyal depths (500-1000 m), during the period of major
subsidence of several thousand metres that formed the Waitemata Basin.

The geology of the northern Hunua Ranges in the vicinity of Hays and Tipakuri
Streams has been mapped by Laws (1931), Healy (1935) and Kear (1959). All three
produced columns to illustrate the basic stratigraphy observed in Hays Stream, as have
Waterhouse (1974) and Hayward and Brook (1984).

Clarke (1905) published the first list of macrofauna collected from Hays Stream (then
called Slippery Stream). He recorded 18 species of molluscs, 4 brachiopods, | echinoderm
and 2 corals. From this locality he described two new species, Flabellum papakurense and
Amussium papakurense. Marwick (in Laws 1931) added a further 7 mollusc species records
from Hays Stream. Squires (1958) added records of three further corals from this locality.

Rec, Auckland Inst. Mus. 29: 113-133 1992

114 EAGLE & HAYWARD

GEOLOGY (Fig. 1)

The northern Hunua Ranges are primarily composed of Permian to Jurassic greywacke
(Waiheke Group, Schofield 1976) capped in places by the eroded remnants of two thin
middle Cenozoic transgressive sequences. The older of these is a late Eocene to early
Oligocene sequence. Only remnant patches of the lower parts of this Te Kuiti Group
sequence (Waikato Coal Measures and Mangakotuku Siltstone) are preserved in this area.
The younger transgressive sequence is of the early Miocene Waitemata Group. It
progressively buried an irregular coastal topography of actively eroding greywacke and
lower Te Kuiti Group rocks.

The basal Waitemata Group rocks (Kawau Subgroup, Hayward and Brook 1984)
consist of c. 10 m of shallow-water sandy conglomerate and sandy limestone (Papakura
Limestone), that pass laterally into c. 8 m of fossiliferous, partly glauconitic, calcareous
sandstone (Tipakuri Sandstone). The best exposed sequence of basal conglomerate and
limestone occurs on the northern bank of Hays Stream (Fig. 1) and the best exposed
sequence of fossiliferous sandstone occurs in Tipakuri Stream.

roads

fault

macrofossiliferous

Hunua Rd cutting

mudstone
sandstone

conglomerate

coal Wk Warkworth Subgroup Waitemata

Grou
limestone K Kawau Subgroup p

c ¢<¢ WCM Waikato Coal Measures

tj ; SS RQ wer Waiheke Group greywacke

Tipakuri
Me OT , Stream
Hays Stream re Tipakuri Stream
section section

Fig. 1. Geological map of the northwestern Hunua Ranges (after Kear 1959 and
Schofield 1976) showing the location of early Miocene stratigraphic columns for Hays
and Tipakuri Streams; n.e. = not exposed.

HAYS & TIPAKURI STREAMS 115

As with other thin, basal Waitemata sequences, these shallow-water sediments pass
rapidly up into deeper water deposits. This transition records a period of substantial
subsidence early in the history of the Waitemata Basin (Ricketts et al. 1989). Insome places,
deep bathyal flysch, transported into the basin from the north-west, directly overlies the
shallow-water sequence and indicates a period with little or no sediment supply during
subsidence. Elsewhere (e.g. Motutapu, Motuihe), 1-20 m of locally-derived sediment
(greywacke lithoclasts and bioclasts) accumulated in upper bathyal depths and document
the episode of rapid subsidence (Ricketts et al. 1989) adjacent to an eroding coastline that
had not yet been drowned. In most parts of the northern Hunua Ranges there is an abrupt
transition from basal shallow-water Waitemata sediments to deep bathyal flysch (Hayward
and Brook 1984, fig. 3), except at Hays Stream. Here, at the downstream end of the Hays
Stream gorge are 30 m of locally-derived sandstone, siltstone and shelly conglomerate
(greywacke and Te Kuiti Group lithoclasts) exposed in the stream-bed and bluffs above and
below Hunua Road. There are no exposed lower or upper contacts for this unit, but the
stratigraphic sequence elsewhere indicates that it overlies the basal, shallow-water Kawau
Subgroup exposed 500 m upstream, and underlies the deep bathyal flysch sequence that
outcrops in the hills within | km to the north.

This paper documents the early Miocene macrofauna and foraminiferal microfaunas
of the Kawau Subgroup at Tipakuri Stream, the microfauna of the Kawau Subgroup at Hays
Stream, and the macrofauna and microfaunas of the locally-derived transitional Waitemata
Group sequence at Hays Stream (Fig. 1).

Fossil record numbers are those of the New Zealand Fossil Record File. All
macrofossils are held in the collections of Auckland Museum and all microfossils are held
by the Institute of Geological and Nuclear Sciences, Lower Hutt.

MACROFAUNA

All collected taxa are listed in Appendix 1. The paleoecological and paleoenvironmental
assessment that follows is largely based on the known ecology of genera living today and
of modern species most closely related to these fossils.

HAYS STREAM FAUNA (R12/f72)

The rich macrofauna from the transitional Waitemata Group sequence collected from
a cutting on the Papakura-Hunua Road above Hays Stream has variable preservation, with
some specimens in near perfect condition and others partly decalcified or broken into many
pieces, probably during pre-burial transport.

The Hays Stream fauna occurs in a bed of muddy, shelly conglomerate and appears
to have been mixed during pre-burial downslope transport in a subaqueous mass flow,
Faunal elements from at least four different communities, or groups of communities, are
identifiable within this single Hays Stream fossil fauna and are discussed separately below,

Intertidal and Shallow Subtidal Mud and Sand Communities, 0-10 m (Fig. 2)

A significant number of fossils in the Hays Stream fauna are inferred to have lived in
or on soft sediment intertidally or to depths no greater than 10 m. Many of these fossils were

116 EAGLE & HAYWARD

ANN
NN

Fig. 2. Schematic drawing of Hays Stream intertidal and shallow subtidal mud and sand

communities (0-10 m). A = Austrovenus n.sp.; Aa = Austrofusus (Neocola) alpha; An

= Arnoglossus novus; Lt = Lutraria trapezoidalis; No = Nucula otamatea; P = Paphies

n.sp.; Pe = Pyrazus consobrinus; Pf = Paracominia finlayi; P1 = Paracominia lignaria;

Th = “Tellina” hesterna; Tr = “Tellina” robini; Zb =Zeafallacia benesulcata; Zp =
Zeacolpus pukeuriensis.

infaunal deposit-feeding bivalves, such as two species of the long siphonate genus
“Tellina’”’, the elongate mactrid Lutraria trapezoidalis, undescribed species of the cockle
Austrovenus and tuatua Paphies, and the small nut shell Nucula otamatea.

The shallow water gastropods may be divided into those that lived in exposed and those
that lived in sheltered conditions. From exposed environments there is the extinct
Austrofusus (Neocola), which presumably lived alongside Zethalia at or below low tide on
a moderately exposed sandy beach. Zeacolpus probably lived in current-swept shallow
subtidal situations. Elements from a sheltered, slightly brackish estuarine environment
include the carnivorous gastropods Zefallacia, Paracominia and Pyrazus. The latter genus
is currently a “mangrove creeper” in tropical areas (Beu et al. 1990). The high-spired,
infaunal gastropod Zeacuminia probably resided in upper intertidal brackish situations, as
it does today. Otoliths tell of the presence of the left-eyed flounder Arnoglossus which
presumably also lived around this environment.

Intertidal and Shallow Subtidal Rock and Coarse Gravel Communities, 0-30 m (Fig. 3)

Intertidal and shallow subtidal rocky reef dwellers in the Hays Stream fauna include
the rock boring bivalve Parapholas aucklandica, the sedentary byssate Pteria oneroaensis
and an unidentified oyster. Rocky reef inhabiting gastropods include the intertidal
camivores Lepsiella and Pagodula waitemataensis, and the very large Sarmaturbo
superbus.

HAYS & TIPAKURI STREAMS 117

Pieces of the hermatypic corals Cyphastrea and Leptastrea are inferred to have
originated from scattered heads growing on a coarse gravelly substrate in shallow water.
Other members of this “subtidal gravel bank” community probably included the ahermatypic
corals Dendrophyllia, Oculina and Flabellum, and the brachiopods Magasella, Notosaria
and Terebratulina. The large, thick-shelled bivalves Glycymerita, Eucrassatella and
Tucetona probably also lived in this shallow gravel community as did the epifaunal
carnivorous gastropods Conus and Conilithes and the large cidarid Phyllacanthus titan,
which is represented in the fossil fauna by its large spines. The ciliary-feeding Crepidula
probably lived in clusters attached to pebbles or large shells and Chama lived cemented by
one valve to any available hard substrate.

Fig. 3. Schematic drawing of Hays Stream intertidal and shallow subtidal rock and
coarse gravel communities (0-30 m). C = Chama n.sp.; Ca = Conus (s.1.) armoricus;
Ce = Cyphastrea cf. chalcidium; Cw = Conolithes wollastoni; Cy = Cyathoceras sp.;
Db = Dendrophyllia boschmai; Ea= Eucrassatella ampla; Fp = Flabellum pavoninum;
Gm = Glycymerita (Manaia) cf. manaiaensis; Lm = Lepsiella maxima; Lt = Leptastrea
cf. transversa; Mn = Magasella neozelandica; Na = Notosaria antipoda; Nv =
Notobalanus vestitus; O = Ostrea sp.; Ov = Oculina cf. virgosa; Pa = Parapholas
aucklandica; Po = Pteria oneroaensis; Pt = Phyllacanthus titan; Pw = Pagodula
waitemataensis; Ss = Sarmaturbo superbus; Ta = Tucetona aucklandica; Ts =
Terebratulina suessi,

118 EAGLE & HAYWARD

Inner to Mid Shelf Silt, Sand and Fine Gravel Communities, 10-100 m (Fig. 4)

Gastropods that are inferred to have lived in this community are abundant and diverse
in the Hays Stream fauna. Apart from the deposit-feeding turritellids Maoricolpus,
Tropicolpus and Zeacolpus, the majority were. carnivorous predators e.g. Alcithoe,
Polinices, Cabestana, Pterynotus, Eumitra, Amalda, Austrotoma and Gemmula.

The deep-burrowing, long-siphoned, suspension feeders Dosinia, Kuia and Panopea
seem to have been the dominant bivalves. Also present are the shallower burrowing
suspension feeders Notocorbula and Caryocorbula and the byssate bivalve Chlamys.
Isognomon possibly lived epifaunally as a cluster of individuals byssally attached to each
other, whereas Lentipecten lived on a sandy substrate and was able to swim off if disturbed.
The infaunal scaphopods Antalis, Fissidentalium and Dentalium were probably all present
in the silt and fine sand of this habitat.

Several different free-living or attached forms of bryozoa and the ahermatypic corals
Flabellum, Truncatoflabellum, Caryophyllia japonica and Oculina may also have lived in
this community. Otoliths suggest the presence in this habitat of the burrowing bandfish
Cepola and the demersal cardinal fish Coelorhynchus. The sand shark Odontaspis is
represented by several fossil teeth.

Fig. 4. Schematic drawing of the Hays Stream inner to mid shelf silt, sand and fine
gravel communities (10-100 m). Ae = Austrotoma excavata; Ap = Antalis pareorensis;
Ar = Amalda (Baryspira) robusta; At = Alcithoe (Alcithoe) turrita; Cf = Chlamys
fischeri; Cm = Cepola cf. macrophthalma; Cn = Caryocorbula nitens; Ct = Cabastana
tetleyi; Db = Dosina (Raina) bensoni; Dm = Dentalium mantelli; E = Epigonus sp.; Ew
= Eumitra waitemataensis; F = Fissidentalium n.sp.; Gk = Gemmula kaiparaensis; Iz
= lsognomon aff. zealandicus; Kv = Kuia vellicata; Lh = Lentipecten hochstetteri; Mw
= Maoricolpus waitemataensis; Nh = Notocorbula humerosa; Oe = Odontapsis
elegans; Ov = Oculina virgosa; Ph= Polinices huttoni; Pk= Pterynotus cf. kaiparaensis;
Po = Polinices oneroaensis; Pw=Panopea worthingtoni; Tg=Ti ropicolpus (Amplicolpus)
gittosinus; Ts = Truncatoflabellum sphenodeum.

HAYS & TIPAKURI STREAMS 119

Outer Shelf to Upper Bathyal Soft Bottom Community, 100-500 m (Fig. 5)

Elements in the Hays Stream fauna that appear to be derived from a quiet, moderately
deep water, soft bottom habitat include abundant shells of the deposit-feeding bivalve
Limopsis, and less common Pseudoportlandica, Neilo, Bartrumia, Lima and Mesopeplum.
Carnivorous gastropods that probably lived at these depths include Nassarius, Chicoreus,
Teremelon, Bathytoma and the opisthobranchs Acteon, Cylichnania and Cylichnina.

Ahermatypic corals that were attached to pebbles or shells or nestled in these soft
sediments at outer shelf or upper bathyal depths probably include Notocyathus pedicellatus,
N. conicus, Trochocyathus papakurensis, Stephanotrochus and the gorgonian Parisis
hamiltoni. The barnacles Smilium, Notobalanus, Tasmanobalanus and Graviscalpellum all
probably lived at these depths (Buckeridge 1983) attached to shells or pebbles that lay on
the sea floor.

Fish otoliths that probably originated at these depths include those of the cosmopolitan
Melanonidarum, an uncommon pelagic cod, and the alfonsino Centroberyx. Another is the
ghost flathead Hoplichthys, which is a deep water benthic fish.

TIPAKURI STREAM FAUNA (S12/f72) (Fig. 6)

The macrofauna from Tipakuri Sandstone (Kawau Subgroup) in the bank of Tipakuri
Stream comes from a5 m thick bed of slightly glauconitic, calcareous sandstone. The fossils
are badly decalcified and many are distorted by post-burial deformation.

The Tipakuri fauna appears to lack the mixing that characterises the Hays Stream
fauna, and is essentially an in-situ fauna from mid to outer shelf depths (50-200 m). It is
dominated by the infaunal suspension feeding bivalves Bartrumia oneroaensis, Spissatella
trailli and Panopea worthingtoni, the epifaunal bivalve Limopsis, and the carnivorous
naticid gastropod Polinices. Other members of this muddy sand community include Lima
colorata, Fissidentalium and rare Struthiolaria lawsi.

RECORDS OF NEW OR RARE TAXA

The Hays Stream fossil fauna contains several mostly shallow water molluscs that
have only been found in one or two other localities, largely because of the rarity of intertidal
and shallow subtidal fossil assemblages in the early Miocene of New Zealand. These
include: Preria oneroaensis, Lutraria trapezoidalis, Dosinia (Raina) bensoni, Sarmaturbo
superbus, Zefallacia benesulcata, Pyrazus consobrinus, Tropicolpus (Amplicolpus)
gittosinus, Cabestana tetleyi, Parapholas aucklandica, Lepsiella intermedia and L. maxima.
A possible new genus of the gastropod family Cerithiidae and a coelenterate of the family
Caryophylliidae, are also recorded. Also present are apparently new species of the bivalves
Neilo, Mesopeplum, Chama, Paphies and Austrovenus, the gastropods Zethalia, Proerato
and Lepsiella and the scaphopod Fissidentalium. A new record for the Waitemata Basin is
the rare serpulid, Sclerostyla ouyenensis, previously recorded by Finlay (1924) from
Miocene strata at Pukeuri and Clifden, Uncommon species found in the Tipakuri Stream
fauna are Struthiolaria lawsi and Bartrumia oneroaensis.

120 EAGLE & HAYWARD

Fig. 5. Schematic drawing of the Hays Stream outer shelf to upper bathyal soft bottom
community (100-500 m). Ao = Acteon oneroaensis; Bb = Bathytoma (Bathytoma)
bartrumi, Ca = Coelorhynchus australis; Ck = Chicoreus (Siratus) komiticus; Cp =
Centroberyx cf. Trachichthodes pulcher; Fp = Flabellum pavoninum; Hm = Hoplichthys
multistriatus; Lc = Limopsis catenata; Lel = Lima colorata; Lz = Limopsis zelandica;
M = Mesopeplum n.sp.; Ma= Melanonidarum aff. Karrerichthys admirabilis; N = Neilo
n.sp.; Np = Notocyathus (Paradeltocyathus) pedicellatus; Ns = Nassarius (Hima)
separabilis, Ph = Parisis hamiltoni; Tg = Tasmanobalanus grantmackiei; Tp =
Trochocyathus (Aplocyathus)papakurensis; Tt= Teremelon tumidor.

Fig. 6. Schematic drawing of the Tipakuri Stream in-situ sandy soft bottom community (50-200 m).

Bo = Bartrumia oneroaensis; F = Fissidentalium n.sp.; Let =Limopsis catenata; Lcl= Lima colorata;

Lz = Limopsis zelandica; Ph = Polinices huttoni; Po = Polinicesoneroaensis; Pw = Panopea
worthingtont; S| = Struthiolaria lawsi; St = Spissatella trailli.

HAYS & TIPAKURI STREAMS 121

MICROFAUNA

Six sediment samples (4 from Hays Stream, 2 from Tipakuri Stream - Fig. 1) were
processed for microfaunal assessment and a quantitative pick of 100 benthic foraminifera
was made from five of these, An assessment of the percentage of planktic foraminifera in
each sample was also made. All taxa have been identified and their counts tabulated in
Appendix 2, Paleoenvironmental assessments are largely based on Hayward (1986).

PLANKTIC FORAMINIFERA

The Hays Stream faunas contain varied percentages of planktics. The lowest sample
(R12/f8) contains a low 10% planktics suggesting neritic waters overhead and probably
inner or mid shelf depths (0-100 m). The two middle, sandstone samples (R12/f9, £7627)
contain 95% and 75% planktics indicating oceanic to marginal oceanic water overhead and
paleodepths in the upper to mid bathyal range (200-2000 m). The highest sample, the matrix
of the richly fossiliferous sandy conglomerate (R12/f14A), contains 40% planktics which
indicates marginal neritic overhead water and suggests an outer shelf to uppermost bathyal
depth range, The planktic fauna in these last three samples is diverse and contains numerous
large specimens, which is also indicative of oceanic, bathyal conditions.

The two Tipakuri Stream faunas contain 25% and 45% planktics indicating marginal
neritic waters usually associated with a mid shelf to uppermost bathyal depth range (50-
400 m),

BENTHIC FORAMINIFERA

Hays Stream - R12/f8

The lowest Hays Stream fauna is dominated by Gaudryina convexa, Cribrorotalia
ornatissimum, Melonis simplex, Cibicides mediocris, C. notocenicus and Elphidium
gibsoni. This assemblage is characteristic of those living in a high energy, coarse sediment,
shallow water environment (inner shelf, 10-50 m). There are no genera present that are
today restricted to deeper water. Indeed many of the rarer taxa are also characteristic of
shallow water (e.g. Amphistegina, Discorbis, Eponides repandus, Haynesina depressula,
Notorotalia powelli, Pileolina, Planoglabratella and Textularia hayi).

Hays Stream - R12/f9

This benthic fauna is sparse and consists ofa mixture of several characteristically inner
shelf taxa (Cribrorotalia, Cibicides notocenicus) and a number of quiet water taxa typical
of outer shelf or upper bathyal depths (Alabamina, Gyroidina, Oridorsalis, Stilostomella
pomuligera).

Hays Stream - R12/f7627

The fauna is dominated by Cibicides mediocris, C. temperatus, C. vortex, Bolivina
reticulata and Trifarina parva. This assemblage is similar to thanatotope B of Hayward and
Buzas (1979) which is characteristic of outer shelf depths (c. 100-200 m). Also present in
f7627 are several typical bathyal or bathyally-restricted taxa (e.g. Stilostomella pomuligera,

122 EAGLE & HAYWARD

Oridorsalis, Osangularia, Planulina). The shallowest known modern record of Osangularia
is 700 m (Hayward and Buzas 1979: 21) anda similar upper depth limit has been determined
from its occurrence in the Miocene of Taranaki (Hayward 1990, table 1). This fauna
therefore indicates that most of the sand was derived from outer shelf depths and was
transported downslope, picking up minor amounts of bathyal sediment en route, before
coming to rest in the deep upper bathyal (c.500-1000 m).

Hays Stream - R12/f14A

The benthic fauna in the matrix of this richly macrofossiliferous conglomerate appears
to be a mixture of two assemblages. It is dominated by robust Amphistegina aucklandica,
a species diagnostic of a high energy, inner shelf environment. Other shallow subtidal taxa
present include Arenodosaria antipoda, Cibicides notocenicus, Dorothia minima and the
larger foraminifer Lepidocyclina orakiensis. Subdominant are Cibicides mediocris, C.
vortex, C, temperatus and Globocassidulina subglobosa - an assemblage typical of finer
sediment at outer shelf depths.

Tipakuri Stream

Both faunas (S12/f16, f18) are dominated by a combination of Nonionella
novozealandica, Cassidulina laevigata, Cibicides mediocris, C. temperatus, C. vortex,
Bolivina mantaensis and B. semitruncata. This assemblage is characteristic of sheltered
muddy sand at mid to outer shelf depths (50-150 m), The lower sample (f16) has
Cribrorotalia ornatissima and Virgulopsis pustulata as additional codominants suggesting
a slightly less sheltered and shallower environment than the higher (f18). These two faunas
are very similar in composition to assemblages in muddy sandstones in the basal Waitemata
sequence on Waiheke Island (pers. obs.).

AGE

The presence of the molluscs Alcithoe turrita (Po-P1), Lutraria trapezoidalis (Po-P1),
[ropicolpus gittosinus (Lw-Po), Austrotoma finlayi (Lw-Po) and Parapholas aucklandica
(Po) inthe Hays Stream fauna (R12/f72) gives an early Miocene (Otaian, Po) age. An Otaian
age is also inferred for the Tipakuri Stream fauna based on the presence of Bartrumia
oneroaensis which is known only from Otaian strata at Oneroa, Waiheke Island and in the
Mount Harris Formation (Beu et al. 1990).

The presence of the foraminifera Catapsydrax dissimilis (Lwh- Po), Haeuslerella
hectori (Lw-Po) and Ehrenbergina marwicki(Po-P1) in the Hays Stream microfaunas (R12/
7627) confirms an Otaian age. The Tipakuri Stream foraminiferal faunas do not provide
such a precise age diagnosis, although they are consistent with a similar age to the Hays
Stream faunas.

PALEOCLIMATE
MACROFAUNA

The macrofossil faunas of Hays and Tipakuri Streams are rich in genera that live today
in warm, subtropical waters at or beyond the northern extreme of the New Zealand region,

HAYS & TIPAKURI STREAMS 123

These include the molluscs Preria, Isognomon, Chama, Eucrassatella, Lutraria, Parapholas,
Pyrazus, Proerato, Polinices (s.s.), Chicoreus (Siratus), Conus (s.\.), Conilithes, Gemmula,
and Bathytoma(s.s.) and the cidarid echinoderm Phyllacanthus (Beuetal. 1990, Fell 1954).
In addition there are extinct genera, some apparently endemic to New Zealand, that by their
affinities also appear to indicate subtropical conditions in New Zealand in the Otaian. These
include Lentipecten, Spissatella, Bartrumia, Kuia, Dosinia (Raina), Sarmaturbo, Zefallacia,
Pareora, Austrofusus (Neocola), Zeacuminia, Austrotoma, and Maudrillia (Beu et al.
1990).

Pieces of two genera of reef coral nave been found in the Hays Stream fauna. No reef
corals currently live around mainland New Zealand. Comparison of the full early Miocene
reef coral fauna of northern New Zealand with the modern ranges of the genera indicates
that sea surface temperatures were 5-7°C warmer than present (Hayward 1977).

MICROFAUNA

The foraminiferal faunas of Hays and Tipakuri Streams include locally abundant
Amphistegina aucklandica and occasional specimens of the larger foraminifera
Lepidocyclina orakiensis, Neither larger foraminifera nor Amphistegina live around New
Zealand today, but both are abundant in shallow environments in the tropical South Pacific
from Norfolk Island northwards. These occurrences support inferences from other fossil
groups that the early Miocene climate of northern New Zealand was marginally tropical
(Hornibrook 1971).

DISCUSSION
TIPAKURI STREAM PALEOENVIRONMENT

Both macro- and microfaunas (S12/f16, f18, £72) appear to be in-situ, unmixed
assemblages that indicate a mid shelf environment deepening to outer shelf depths during
the period of accumulation.

HAYS STREAM PALEOENVIRONMENT

The microfossil faunas from the Miocene sequence in Hays Stream support previous
interpretations of a transgressive sequence. The basal conglomerate and limestone (Kawau
Subgroup, R12/f8) accumulated at inner shelf depths in a high energy environment. The
area had subsided to deep upper bathyal (500-1000 m) by the time the transitional
Waitemata Group sequence was accumulating. The micro- and macrofaunas (R12/f9,
f14A, £7627, £72) indicate that much of the sand and fine gravel component that
accumulated in this transitional sequence: was derived from various environments at
bathyal, shelf and intertidal depths (0-500 m). The coarser clasts can be confidently
identified as being derived locally from eroding Waiheke Group greywacke and Waikato
Coal Measures.

Thus, during accumulation of the transitional Waitemata sequence, there was still a
landmass or island in the vicinity, with an eroding rocky shoreline interspersed with sandy
beaches and a sheltered, mangrove-lined bay or inlet. Offshore there were areas at mid and
outer shelf depths accumulating thin sedimentary sequences and supporting rich benthic

124 EAGLE & HAYWARD

communities. The seafloor probably dropped fairly steeply down to bathyal depths and
large amounts of the locally-derived shallow water sediment and fauna were transported
downslope by subaqueous mass flows.

Further subsidence drowned these nearby land areas before accumulation of the
overlying Waitemata Group flysch. Paleocurrent and heavy mineral evidence indicate that
the flysch has been transported into the depths of the Waitemata Basin as turbidity currents
sourced from the Kaipara area, 50 km to the north-west (Hayward and Smale 1992),

Acknowledgements. We wish to thank Nick de Carteret and Chris Hollis for assistance in the field.
We are also grateful to John Buckeridge (Carrington Polytechnic), Jack Grant-Mackie (University
of Auckland) and Fred Brook (Department of Conservation, Whangarei) for assistance with
identification of some of the fossil barnacles, molluscs and corals, respectively. Graeme Gibson
(University of Auckland) kindly gave us access to the Hays Stream microfaunal sample R12/f7627,
which he and his 1970 micropaleontology class collected and processed. Some of the fossil
illustrations in Figs. 2-6 are adapted from drawings by Ron Brazier (Beu et al. 1990) and the late Jack
Marwick (Fleming 1966). We appreciate the helpful comments of Anthony Wright, Brett Stephenson,
Chris Hollis (all Auckland Institute and Museum), Jack Grant-Mackie and Fred Brook, who kindly
read the draft manuscript.

REFERENCES

Beu, A.G., P.A. MAxweLt and R.C. BRAZIER
1990 Cenozoic Mollusca of New Zealand. New Zealand Geological Survey Paleontological
Bulletin 58.

BuckerIbGE, J.S.
1983 Fossil barnacles (Cirripedia: Thoracica) of New Zealand and Australia. New Zealand
Geological Survey Paleontological Bulletin 50.

Crark, E. de C,
1905 The fossils of the Waitemata and Papakura Series. Transactions of the New Zealand
Institute 37: 413-421.

Dawson, E.W.
1990 The Cenozoic Brachiopoda of New Zealand: A commentary, reference list, and
bibliography. New Zealand Oceanographic Institute Miscellaneous Publication 103.

FELL, H.B.
1954 Tertiary and Recent Echinoidea of New Zealand. New Zealand Geological Survey
Paleontological Bulletin 23.

FINLAY, H.J.
1924 Three fossil annelids new to New Zealand. Transactions of the New Zealand
Institute 55: 448-449.

FLEMING, C.A,
1966 Marwick’s illustrations of New Zealand shells, with a checklist of New Zealand
Cenozoic Mollusca. Department of Scientific and Industrial Research Bulletin 173.

HAYS & TIPAKURI STREAMS 125

GRENFELL, H.R.
1984 Early Miocene teleost otoliths from Parengarenga Harbour, New Zealand. New
Zealand Journal of Geology and Geophysics 27: 51-96.

Haywarp, B.W.
1977 Lower Miocene corals from the Waitakere Ranges, North Auckland, New Zealand.
Journal of the Royal Society of New Zealand 7: 99-111.
1986 A guide to paleoenvironmental assessment using New Zealand Cenozoic foraminiferal
faunas. New Zealand Geological Survey Report Pal 109.
1990 Use of foraminiferal data in analysis of Taranaki Basin, New Zealand. Journal of
Foraminiferal Research 20: 71-83.

Haywarp, B.W. and F.J. Brook
1984 —_Lithostratigraphy of the basal Waitemata Group, Kawau Subgroup (new), Auckland,
New Zealand. New Zealand Journal of Geology and Geophysics 27: 101-123.

Haywarp, B.W. and M.A. Buzas
1979 Taxonomy and paleoecology of early Miocene benthic foraminifera of northern
New Zealand and the north Tasman Sea. Smithsonian Contribution to Paleobiology 36.

Haywarp, B.W. and D. SMALE
1992 Heavy minerals and the provenance history of Waitemata Basin sediments (early
Miocene, Northland, New Zealand). New Zealand Journal of Geology and
Geophysics 35: 223-242.

HEALY, J.
1935 Geology of the Hunua-Ramarama area. Unpublished M.Sc. thesis, Geology Department,
University of Auckland.

Hornisrook, N.deB.
1971 New Zealand Tertiary climate. Report New Zealand Geological Survey 47.

HorniBrook, N.deB., C.P. Strone and R.C. Brazier
1989 Manual of New Zealand Permian to Pleistocene foraminiferal biostratigraphy.
New Zealand Geological Survey Paleontological Bulletin 56.

Kear, D.
1959 Drury coalfield, Auckland. New Zealand Journal of Geology and Geophysics 2:
846-869.
Laws, C.R.

1931 Geology of the Papakura-Hunua district, Franklin County, Auckland. Transactions
of the New Zealand Institute 62: 37-66.

Ricketts, B.W., P.F. BALLANcE, B.W. Haywarp and W. Mayer
1989 Basal Waitemata Group lithofacies: rapid subsidence in an early Miocene interarc basin,
New Zealand. Sedimentology 36: 559-580.

SCHOFIELD, J.C.
1976 Sheet N48 Mangatawhiri (Ist ed.), Geological Map of New Zealand 1:63 360.
DSIR, Wellington.

126 EAGLE & HAYWARD

SCHWARZHANS, W.
1984 Fish otoliths from the New Zealand Tertiary. New Zealand Geological Survey

Report 113.

Squires, D.F.
1958 The Cretaceous and Tertiary corals of New Zealand. New Zealand Geological Survey

Paleontological Bulletin 29.
1962 Additional Cretaceous and Tertiary corals from New Zealand. Transactions of the
Royal Society of New Zealand 1: 133-150.

STINTON, F.C.
1957 Teleostean otoliths from the Tertiary of New Zealand. Transactions of the Royal

Society of New Zealand 84: 513-517.

WATERHOUSE, B.C.
1974 Type and standard localities: Papakura Limestone. New Zealand Journal of Geology and

Geophysics 17: 487-490.

M. EAGLE and B.W. HAYWARD, Auckland Institute and Museum, Private Bag 92018,
Auckland.

HAYS & TIPAKURI STREAMS 127

APPENDIX 1. Systematic list of macrofauna tabulated from Hays Stream (H) and Tipakuri
Stream (T) localities. Taxonomy follows Beu et al. (1990) for Mollusca; Stinton (1957),
Grenfell (1984) and Schwarzhans (1984) for otoliths; Dawson (1990) for brachiopods;
Buckeridge (1983) for barnacles; Squires (1958, 1962) for coelenterates.

MOLLUSCA
BIVALVIA
NUCULIDAE Nucula otamatea Laws, 1939 H
NUCULANIDAE Pseudoportlandica cf. solenelloides (Marshall, 1919) H
Ledella pakurangiensis Laws, 1941 H
MALLETIIDAE Neilo n.sp. H
LIMOPSIDAE Limopsis catenata Suter, 1917 H
Limopsis zealandica Hutton, 1873 H
Limopsis sp. sa
GLYCYMERIDIDAE Glycymerita (Manaia) cf. manaiaensis (Marwick, 1923) H
Tucetona aucklandica (Powell, 1938) H,T
PTERIIDAE Pteria oneroaensis (Powell & Bartrum, 1929) H
ISOGNOMONIDAE Isognomon aff. zealandicus (Hutton in Suter, 1917) H
PECTINIDAE Chlamys fischeri (Zittel, 1864) H
Lentipecten hochstetteri (Zittel, 1864) H
Mesopeplum costatostriatum (Marshall, 1918) H
Mesopeplum n.sp. H
Mesopeplum sp. H
LIMADAE Lima colorata Hutton, 1873 H,T
OSTREIDAE Ostrea sp. H
CHAMIDAE Chama _ n.sp. H
CARDITIDAE “Cyclocardia” awamoensis (Harris, 1897) H
Glyptoactis (Fasciculicardia) subintermedia (Suter, 1917) H
CRASSATELLIDAE Eucrassatella ampla (Zittel, 1864) H,T
Spissatella tralli (Hutton, 1873) T
MACTRIDAE Lutraria trapezoidalis Powell & Bartrum, 1929 H
MESODESMATIDAE Paphies n.sp. H
TELLINIDAE “Tellina” hesterna (Powell & Bartrum, 1929) H
“Tellina” robini (Finlay, 1924) H
Bartrumia oneroaensis (Powell & Bartrum, 1929) H,T
VENERIDAE Kuia vellicata (Hutton, 1873) H
Austrovenus sp. H
Dosinia (Raina) bensoni Marwick, 1927 H
Dosinia lambata (Gould, 1850) T
CORBULIDAE Caryocorbula nitens (Marshall, 1919) H
Notocorbula humerosa (Hutton, 1885) H
Notocorbula cf. pumila (Hutton, 1885) H
HIATELLIDAE Panopea worthingtoni Hutton, 1873 H
PHOLADIDAE Parapholas aucklandica Powell, 1938 H

128 EAGLE & HAYWARD

GASTROPODA

TROCHIDAE Zethalia n.sp.
TURBINIDAE Sarmaturbo superbus (Zittel, 1864)
CERITHIIDAE Zefallacia benesulcata Powell & Bartrum, 1929

n.gen. & n.sp.
PAREORIDAE Pareora striolata (Hutton, 1885)
POTAMIDIDAE Pyrazus consobrinus Powell & Bartrum, 1929
TURRITELLIDAE Maoricolpus waitemataensis (Powell & Bartrum, 1929)

Tropicolpus (Amplicolpus) gittosinus (Powell & Bartrum, 1929)
Zeacolpus pukeuriensis Marwick, 1934
Zeacolpus tetleyi (Powell & Bartrum, 1929)
STRUTHIOLARIIDAE — Struthiolaria lawsi Powell & Bartrum, 1929
CALYPTRAEIDAE Crepidula (Maoricrypta) aff. opuraensis Powell & Bartrum, 1929
TRIVIIDAE Proerato n.sp.
NATICIDAE Polinices huttoni lhering, 1907
Polinices oneroaensis Powell & Bartrum, 1929
Polinices sp.
RANELLIDAE Cabestana tetleyi Powell & Bartrum, 1929
BUCCINIDAE Austrofusus (Neocola) alpha (Finlay, 1926)
Austrofusus (Neocola) oneroaensis Powell & Bartrum, 1929
Paracominia finlayi (Powell & Bartrum, 1929)
Paracominia lignaria (Powell & Bartrum, 1929)
NASSARIIDAE Nassarius (Hima) separabilis (Laws, 1939)
MURICIDAE Chicoreus (Siratus) komiticus (Suter, 1917)
Lepsiella intermedia Powell & Bartrum, 1929
Lepsiella maxima Powell & Bartrum, 1929
Lepsiella n.sp.
Pagodula waitemataensis (Powell & Bartrum, 1929)
Pterynotus cf. kaiparaensis Fleming, 1962
MITRIDAE Eumitra nitens (Marshall, 1918)
Eumitra waitemataensis (Powell & Bartrum, 1929)
VOLUTOMITRIDAE Proximitra partinoda Finlay, 1930
OLIVIDAE Amalda (Baryspira) platycephalia (Powell & Bartrum, 1929)
Amalda (Baryspira) robusta (Marwick, 1924)
Amalda (Spinaspira) stortha (Olson, 1956)

VOLUTIDAE Alcithoe (Alcithoe) turrita (Suter, 1917)
Teremelon tumidor (Finlay, 1926)
CONIDAE Conilithes wollastoni Maxwell, 1978

Conus (sensu lato) thorae Finlay, 1926

Conus (sensu lato) armoricus (Suter, 1917)
TEREBRIDAE Zeacuminia cf. orycta (Suter, 1917)
TURRIDAE Anacithara clifdenica Powell, 1942

Austrotoma excavata (Suter, 1917)

Austrotoma minor (Finlay, 1924)

Bathytoma (Bathytoma) bartrumi Laws, 1939

Gemmula kaiparaensis (Marshall, 1918)

Maudrillia supralaevis Powell, 1942

Tomopleura (Maoritomella) sola (Powell, 1942)

a |

cage ogre ets ors ora asco eases tie oie bgasie sa sie omer mon sesiosesiiso ges.) o nec nano ng ae on ge pao gle Seo eeckiae ters ssn gc nese snes be ile ore cia se,

HAYS & TIPAKURI STREAMS 129

ACTEONIDAE Acteon oneroaensis Powell & Bartrum, 1929

CYLICHNIDAE Cylichnania bartrumi Marwick, 1931
Cylichnina ennucleata Powell & Bartrum, 1929

PYRAMIDELLIDAE Chemnitzia brevisutura Laws, 1937

acts

SCAPHOPODA
DENTALIIDAE Antalis pareorensis (Pilsbry & Sharp, 1897) H
Dentalium mantelli Zittel, 1864 H
Fissidentalium n.sp. H
LAEVIDENTALIIDAE Laevidentalium waihoraense Emerson, 1954 H

BRYOZOA
gen. & spp. indet. H

BRACHIOPODA
CANCELLOTHYRIDIDAE
Terebratulina suessi (Hutton, 1873)
gen. & sp. indet. —
DALLINIDAE Magasella neozelandica (von Ihering, 1903)
HEMITHYRIDIDAE Notosaria antipoda (Thomson, 1918)

2 ome Oae oe 2)

POLYCHAETA
Sclerostyla ouyenensis (Chapman, 1913)
Spirorbis sp.
gen. & sp. indet.

xno

ECHINOIDEA
CIDARIDAE Phyllacanthus titan Fell, 1954 H

CIRRIPEDIA
ARCHAEOBALANIDAE Notobalanus vestitus (Darwin, 1854)
Tasmanobalanus grantmackiei Buckeridge, 1983
SCALPELLIDAE Graviscalpellum ungulatum (Withers, 1913)
?Smilium subplanum (Withers, 1924)

jo ogne ae oes =)

COELENTERATA
CARYOPHYLLIIDAE Caryophyllia japonica Marenzeller, 1888
Cyathoceras sp.
Notocyathus conicus (Alcock, 1902)
Notocyathus (Paradeltocyathus) pedicellatus Tenison-Woods, 1880
Sphenotrochus sp.
Trochocyathus (Aplocyathus) papakurensis (Clarke, 1905)
n.gen. & n.sp.
DENDROPHYLLIIDAE Balanophyllia alta Tenison-Woods, 1880
Dendrophyllia boschmai van der Horst, 1926
FAVIIDAE Cyphastrea cf. chalcidum (Forskaal, 1775)
Leptastrea cf. transversa Klunzinger, 1879

ZSeertretetete eee

130 EAGLE & HAYWARD

FLABELLIDAE

MELITODIDAE
OCULINIDAE

ODONTASPIDIDAE

APOGONIDAE
BERYCIDAE
BOTHIDAE
CARAPIDAE
CEPOLIDAE
HOPLICHTHYIDAE
MACROURIDAE

MELANONIDAE

Flabellum lamellulosum Alcock, 1902

Flabellum pavoninum Lesson, 1831

Tortoflabellum marwicki Squires, 1962
Truncatoflabellum sphenodeum (Tenison-Woods, 1880)
Truncatoflabellum sp.

Parisis hamiltoni (Thompson, 1908)

Oculina virgosa Squires, 1958

Oculina cf. virgosa Squires, 1958

CHONDRICHTHYES
Odontaspis elegans Agassiz, 1843

TELEOSTS
gen. & spp. indet. (teeth)
Epigonus sp.

Centroberyx cf. Trachichthodes pulcher Schwarzhans, 1980

Arnoglossus novus Schwarzhans, 1980
Carapus sp.

Cepola cf. macrophthalma Linnaeus, 1758
Hoplichthys multistriatus Grenfell, 1984
Coelorhynchus australis (Richardson, 1839)
Coelorhynchus cf. C. toulai (Schubert, 1905)

Melanonidarum aff. Karrerichthys admirabilis Schwarzhans, 1980

coe eee

H
H
H
H
H
H
H
H
H
H

HAYS & TIPAKURI STREAMS 131

APPENDIX 2. List of foraminifera obtained from Hays and Tipakuri Stream samples.
Numbers are abundances (%) in each sample from picks of 100 benthic foraminifera. * =
present but not recorded in quantitative pick. Taxonomy follows Hayward and Buzas
(1979) and Hornibrook et al. (1989).

1 R12/f8 Hays Stream inner shelf

2 R12/f9 Hays Stream mixed inner shelf-upper bathyal
3. R12/f7627 Hays Stream mixed outer shelf-upper bathyal
4 RI12/fl4A Hays Stream mixed inner-outer shelf

5 §12/f16 Tipakuri Stream mid-outer shelf

6 §12/f18 Tipakuri Stream mid-outer shelf

Alabamina tenuimarginata
Amphistegina aucklandica 4 58
Anomalinoides fasciatus ] Ys 2 4
Anomalinoides macraglabra ] 2
Arenodosaria antipoda 3
Astrononion parki l
Bolivina acerosa l
Bolivina arta 3

Bolivina cf. acerosa 2
Bolivina mantaensis l 11
Bolivina plicatella mera l
Bolivina reticulata 17 ]
Bolivina semitruncata 4 3
Bolivinopsis cubensis 3

Buccella lotella ]
Bueningia creeki l ]

Cancris lateralis 2 2
Cassidulina laevigata 1 11 10
Cassidulina margareta 6
Catapsydrax dissimilis " *

Cibicides brevoralis

Cibicides lobatulus l

Cibicides mediocris 10 20 13 16 6
Cibicides notocenicus 8 *
Cibicides novozelandicus

Cibicides perforatus l *
Cibicides refulgens Nt
Cibicides temperatus 4 13 2 3 l

NO — — —& bO

Ne WD

132 EAGLE & HAYWARD

Cibicides vortex
Cribrorotalia ornatissimum
Discorbis balcombensis
Dorothia minima
Dyocibicides sp.
Ehrenbergina marwicki
Elphidium advenum
Elphidium gibsoni
Elphidium kanoum
Elphidium pseudoinflatum
Epistominella cassidulinoides
Epistominella iota
Eponides repandus
Fissurina laevigata
Fissurina marginata
Florilus stachei

Gaudryina convexa
Gavelinella zealandica
Globigerina ciperoensis
Globigerina falconensis
Globigerina woodi
Globigerina woodi connecta
Globocassidulina subglobosa
Globoquadrina dehiscens
Globorotalia nana
Globorotalia semivera
Guttulina problema
Gyroidina subzelandica
Gyroidina zelandica
Hanzawaia bertheloti
Hanzawaia cf. complanata
Hanzawaia laurisae
Hanzawaia stachei
Haeuslerella hectori
Haynesina depressula
Kolesnikovella australis
Lenticulina mamilligera
Lenticulina nitida

Lepidocyclina orakiensis orakiensis

Melonis simplex
Nodosaria filiformis
Nonion cassidulinoides

e+ + — * *

~~ _—_— i —_—

—_— § — —

— Wh * *

No —-| — —

HAYS & TIPAKURI STREAMS 133

Nonionella novozealandica 2 2 7 36
Notorotalia powelli 4 * ps
Oridorsalis umbonatus *

Osangularia culter 1

Pileolina patelliformis l
Pileolina radiata 1
Planoglabratella semiopercularis 1
Planulina crassa

Quinqueloculina seminula 1
Rectuvigerina rerensis

Semivulvulina capitata

Siphonina australis

Siphouvigerina proboscidea 1
Stilostomella fijiensis ]
Stilostomella pomuligera “4 it l |
Stilostomella verneuilii

Textularia hayi 2

Trifarina sp. 2
Trifarina costornata pi

Trifarina esuriens 1
Trifarina parva 3 2
Uvigerina picki
Vaginulinopsis recta
Virgulopsis pustulata 4

Nn
—

Percentage planktic foraminifera WW 6h 75 40 45 25

i
*

Road

ECOLOGY AND MORPHOLOGICAL VARIATION IN
PLACOSTYLUS BOLLONSIT (GASTROPODA:
BULIMULIDAE) AT THREE KINGS ISLANDS,
NEW ZEALAND

F.J. Brook* AND C.M. LAURENSON**

* DEPARTMENT OF CONSERVATION
** WHANGAREI

Abstract. Five extant and one extinct populations of the local endemic landsnail
Placostylus bollonsi Suter are known from Three Kings Islands. Morphometric data
show that there are consistent differences between the populations in shell characters,
and historical information indicates that those morphological differences are probably
largely hereditary. The populations on West and North East Islands are estimated to
comprise at least 100 and 200 individuals respectively. The extant populations on Great
Island were reduced to very low numbers earlier this century after habitat destruction
by feral goats, but have increased markedly in size and range since 1946 in response to
regeneration of broadleaf vegetation following the removal of goats from the island. In
March 1991 it was estimated that the P. b. bollonsi population covered an area of about
1.69 ha and numbered at least 210 individuals, the P. b. caperatus population covered
an area of about 0.49 ha and numbered at least 130 individuals, and the P. b. arbutus
population covered an area of about 2.7 ha and comprised at least 360 individuals.

Adult P. bollonsi snails are believed to lay small numbers of relatively large eggs
up to 18 mm long in spring and/or early summer, with most eggs hatching by late
summer. Size frequency data indicate the possibility that juvenile snails increase in
length by about 25-30 mm/year, and attain adult size at about 3 years old. Rates of
turnover of adults in populations are not known, but the average age of adult snails is
likely to be at least 8-10 years given the very low proportions of juvenile size classes
in populations. Size frequency data for empty shells suggest that at least 60% of snails
fail to attain adult size, and indicate that relative mortality rates of hatchlings (i.e. shells
<20 mm high) are significantly higher than those of larger juveniles. Mortality of P.
bollonsi is inferred to occur predominantly as a result of dessication or old age; there
is no evidence that predation is a significant cause of mortality.

The principal factors determining the distribution of P. bollonsi snails within
colonies appear to be the presence of broadleaf food plants, and sheltered microhabitats
in broadleaf litter or under groundcover plants. P. bollonsi occurs at densities of up to
2-6 snails/m? in local patches, but overall mean densities of populations determined
from 25 m? quadrats range from 0.15-0,35 snails/m?.

Rec. Auckland Inst. Mus. 29: 135-166 1992

136 BROOK & LAURENSON

Three Kings Islands, 60 km NW of Cape Reinga, are a WSW-trending island chain
comprising one large island of 407 ha (Great Island), three smaller islands of 10-40 ha each
(North East, West and South West Islands), and several islets and rocks up to 5 ha (Figs.
1,2). Terrestrial habitats at Three Kings Islands support a large number of endemic taxa,
including plants (Oliver 1948; Baylis 1951a; Cranwell 1962), earthworms (Lee 1959),
molluscs (Powell 1948, 1951; Climo 1973), arthropods (e.g. Salmon 1948a, 1948b, 1948c:
Woodward 1950, 1954; Forster 1954; Given 1954; Marples 1956; Giles 1958; Holloway
1963), a skink (McCann 1955), and a subspecies of bellbird (Falla 1948). This report is
concerned with one of the Three Kings Islands endemics, the large landsnail Placostylus
bollonsi. Five geographically isolated extant populations of P. bollonsi are present at Three
Kings Islands, comprising one on each of West and North East Islands, and three on Great
Island. There is also an extinct population of the species known from near Hapuka Point,

Great Island.
Three Kings Islands

maya

North East Island

So
<
le J :

Uke
Crater Heag %; Tiss
c North West ca 120
at: ae
Bay : xe Bert
'
_ South East Bay Three Kings
ia aioe

Great Island Tasman Bay

4 189

FOLD,
Bre <a” Wo
ss

North Island

South West Island

Princes Islands

ot oe
a
VW, West Island

Fig. 1. Geography of Three Kings Islands (Manawa Tawhi), with asterisks denoting

locations of Placostylus bollonsi Suter populations. Spot heights of islands are given

in metres. Areas on Great Island enclosed in boxes are shown enlarged in Figs. 10
and 11.

PLACOSTYLUS 137

7,
°%
Middlesex
00 Bank

ENS
OS 500!

Three Kings \%

AY
| Ae
Bank Be) Ks

Cape Reinga North Cape

Parengarenga
Harbour

Rangaunu — Moturoa Islands
Bay ‘s,

: ( Doubtless
Bay

Harbour

6 Ahipara

Fig. 2. Location and bathymetric setting of Three Kings Islands, northern New Zealand.

This study was undertaken with three aims directed at assessing the conservation status
of populations of P. bollonsi. The first aim was to obtain basic information on the
population structure and life history of the species. The second was to assess changes in
the size and range of P. bollonsi populations on Great Island, since censuses carried out in
1945-46 and 1970 by A.W.B. Powell and F.M. Climo respectively (Powell 1948, Climo
1973). The third aim was to obtain morphometric information on shell shapes in
populations of P. bollonsi collected in 1946-51 and 1982-83, to assess the nature of spatial
and temporal morphological variation within and between populations. The purpose of
carrying out the morphometric study was firstly, to provide a reassessment of the
qualitatively determined morphological groupings of Powell (1948, 1951), and secondly,
to assess the relative influences of environmental induction and genetic selection on shell
character polymorphism within P. bollonsi.

138 BROOK & LAURENSON

Physical Setting

Islands of the Three Kings group are topographically rugged, having maximum
elevations of about 90-295 m, and are mostly bounded by precipitous cliffs. The only
beaches are cobble-boulder beaches in North West and South East Bays on Great Island,
and that island is the only one large enough to have permanent streams. No meteorological
data are available for Three Kings Islands, the nearest recording station being at Cape
Reinga on the adjacent mainland coast (Fig. 2). Geographic proximity, and comparable
elevation, topography and oceanic influence suggest that the two areas probably have
similar climates, although that at Three Kings Islands is likely to be slightly milder and
drier. New Zealand Meteorological Service records from Cape Reinga for the period 1919-
1980 indicate that mean monthly rainfall was lowest for January (58 mm) and highest for
June and July (118-120 mm). For the period 1951- 1980, monthly averages of daily
maximum and minimum temperatures were highest for February (22.1 °C and 16.4 °C
respectively), and lowest for July (14.3 °C and 9.9 °C respectively).

Geological History

Three Kings Islands and at least part of the submarine plateau on which they are located
(i.e. Three Kings Bank, Fig. 2) are formed of indurated, highly deformed and altered marine
sedimentary and volcanic rocks of probable early Cretaceous age (Hayward & Moore 1987;
Brook 1989). The Cretaceous to early Cenozoic history of northern New Zealand is poorly
known, butit is possible that marine conditions persisted in the vicinity of present day Three
Kings Islands for much or all of that time. The present structural configuration of the
northernmost New Zealand area is inferred to have developed largely during the mid
Cenozoic. In early Miocene time (c. 15-20 Ma), an antiformal ridge bounded to the NE by
a wrench fault zone is believed to have formed a subaerial peninsula extending from the
present-day Kaitaia and Doubtless Bay areas, to at least as far NW as the Three Kings
Islands area (Brook & Thrasher 1991). This is probably the only period in the mid to late
Cenozoic during which a continuous land connection existed between Northland and the
Three Kings Islands area. By mid-late Miocene time that ridge is inferred to have
foundered, but small land areas probably persisted as islands in the present-day Reinga-
North Cape and Three Kings Islands areas. Continued movement along the NW-trending
wrench fault zone during this period produced a steep continental slope off northeastern
Northland, and led to formation of Three Kings Trough, North Maria Ridge and King Bank,
and Karetu Trough (Fig. 2). In late Miocene time (i.e. prior to 8 Ma) much of northern
Northland was uplifted by a few hundred metres as indicated by the presence of high level
marine-cut terraces (Brook & Thrasher 1991). Erosional terrace remnants at 70-110 m
elevation on Three Kings Islands were probably also cut during this period (cf. Hayward
& Moore 1987), and point to an increase in land area on Three Kings Bank concomitant
with uplift. There is no evidence that major tectonism has occurred on or adjacent to
northernmost New Zealand since Pliocene time and topographic features such as Three
Kings Bank, Karetu Trough and a broad shelf north and west of Reinga are inferred to have
existed in more-or-less their present form since at least that time (Brook & Thrasher 1991).

Cyclical global climate changes and associated glacio-eustatic sealevel fluctuations
had major effects on the distribution of land areas in Pliocene and Pleistocene time.
Lowered sealevels during glacial periods would have caused significant increases in land
areas in northernmost New Zealand and on Three Kings Bank, and islands would have been

PLACOSTYLUS 139

present at times on Middlesex and King banks. The depth of Karetu Trough ensured that
Three Kings Islands would have remained geographically isolated from mainland Northland
during Pliocene-Pleistocene time, but the intervening marine strait would have narrowed
to as little as 10 km wide during some Pleistocene glacial periods.

During the peak of the ultimate glaciation from about 18-20 Ka when sealevel was
about 120 m lower than at present, all islands in the Three Kings group would have been
part of a single landmass of about 120 km? in area. By extrapolating from present-day
bathymetry using sealevel curves of Carter et a/. (1983) and Gibb (1986), it is inferred that
the islands would have remained connected during the subsequent post-glacial sealevel rise
until about 11-14 Ka. The latest Pleistocene to Holocene geological history of the Three
Kings area has thus been one of fragmentation and reduction of terrestrial habitat areas, and
associated profund climatic changes, during the transition from glacial to interglacial
conditions (cf. Stevens et al. 1988).

Anthropogenic Modification

A population of Maori lived permanently on Three Kings Islands for one or more
periods in prehistoric time (i.e. pre 1800) as indicated by the presence of numerous
cultivation and dwelling sites on Great and North East Islands, and an area of cultivation
sites on South West Island (Hayward 1987). Anthropogenic terraces, midden and a burial
site are present on West Island, but there is no evidence that that island was ever occupied
for long (Baylis 1958; Hayward 1987). Early European records suggest that Three Kings
Islands were probably inhabited by Maori for much of the 17th and 18th centuries (Baylis
1948; Hayward 1987). Abel Tasman recorded the presence of cultivated areas and at least
30-35 Maori on Great Island in January 1643, and noted an apparent absence of trees on
the island at that time (Heeres 1898). Marion du Fresne found that Great Island was
occupied by Maori in 1772 and described it as being grassy with groves of bushes and having
a barren appearance (Roth 1891). D’Entrecasteaux recorded Maori occupation of North
East Island in 1793 (Labillardiére 1799). There was apparently sporadic occupation of
Three Kings Islands during the early part of the 19th century, with the last long term Maori
occupation occurring on Great Island between the 1830s and about 1840 (Puckey 1836;
Cheeseman 1888; McNab 1908; Baylis 1948).

Available information suggests that all of the larger islands in the Three Kings group
had a cover of coastal broadleaf forest before human occupation, but that almost all original
forest on Great and North East Islands, and probably also much or all of that on South West
Island, was cleared by prehistoric Maori in the process of preparing cultivation sites or
obtaining fuel (Baylis 1948, 1958; Hayward 1987). Patchwork forest regeneration may
have occurred if areas were temporarily retired from cultivation because of nutrient
depletion (cf. Leach 1984), but original forest remnants probably only persisted in areas that
were too steep or rocky for cultivation. Maori apparently kept pigs and goats on Great Island
during the early part of the 19th century (Baylis 1948), and feeding and trampling by those
animals would almost certainly have had ari effect on the indigenous fauna and flora.

There has been no long term occupation of Three Kings Islands since departure of the
last Maori population from Great Island at about 1840, but small groups of people have
subsequently visited the islands. Parts of Great Island were fired in the early part of the 20th
century (Baylis 1948; Doak 1969), but the greatest modification of that island since the end

140 BRooK & LAURENSON

of Maori occupation occurred as a result of browsing and trampling by feral goats. There
is no direct record as to whether or not any goats were left behind by departing Maori.
Botanical surveys carried out on Great Island in 1887 and 1889 (Cheeseman 1888, 1891)
record kanuka (Kunzea ericoides) and manuka (Leptospermum scoparium) as the dominant
vegetation cover, but with common admixed broadleaf shrubs and small trees. The
presence of abundant plants of highly palatable broadleaf and herb species on the island at
that time suggests that goats were probably absent (Baylis 1948, 1986). In 1889 4 goats
were liberated on Great Island and a further 2 goats were landed on South West Island to
provide a source of food for possible future castaways (Baylis 1948). The goats on South
West Island apparently failed to establish (Buddle 1948). However, those on Great Island
multiplied rapidly, becoming plentiful by 1903 (D.G. Matthew, pers. comm. in Baylis
1948), and numbering about 300 in 1928 (Fraser 1929). The goat population on Great Island
was exterminated in 1946, with a total of 393 individuals being killed (Turbott 1948).
Between 1889 and 1946, goats on Great Island had the effect of maintaining a kanuka-
dominated subclimax vegetation cover by preventing broadleaf regeneration, and they also
caused the extreme reduction and local extinction of several plant species (Baylis 1948,
Turbott 1948). Scattered broadleaf trees and a few small remnant broadleaf groves survived
locally on steep and/or rocky areas on Great Island, and some shrub and herb species were
restricted to sites on coastal cliffs (Baylis 1948). Regeneration of most of the palatable tree,
shrub, herb and fern species was rapid and widespread following the removal of goats
(Baylis 1948, 1951b; Holdsworth 1951; Holdsworth & Baylis 1967; Cameron et al. 1987),
but kanuka has persisted as the dominant canopy cover over much of the island.

In contrast to Great Island, broadleaf forest dominated by large pukanui (Meryta
sinclairii) and pohutukawa (Metrosideros excelsa) trees was established over much of the
formerly cultivated top of North East Island by 1947-55, with kanuka being dominant
locally in forest and shrubland near the summit, and common in fringing shrubland on the
upper slopes of the island (Buddle 1948, Baylis 1958). The vegetation cover of North East
Island, including a small patch of bracken thought to be of anthropogenic origin (Baylis
1958), has apparently subsequently changed little from that described in 1947-55.

The vegetation cover on South West Island in the late 19th century consisted of
pukanui-dominated forest on the upper slopes, and shrublands with common manuka and
kanuka over much of the rest of the island (Cheeseman 1891). Pukanui-dominated forest
subsequently spread to cover just about all except the coastal cliffs, leading to the
elimination of manuka and virtual disappearance of kanuka from the island by 1950-52
(Baylis 1958). |

Differences in the relative proportions of kanuka/manuka-dominated and broadleaf-
dominated forest cover, and observed temporal displacement of kanuka and manuka by
pukanui, led Baylis (1958) to conclude that Great, North East and South West Islands
represent a series of stages in the regeneration of indigenous broadleaf communities. This
suggests the possibility that Maori occupation may have persisted longer on North East
Island than it did on South West Island (Baylis 1958), assuming that rates of vegetation
succession on the two islands are similar.

West Island is thought to be the only large island in the Three Kings group on which
indigenous broadleaf forest and shrubland communities were not extensively modified or
largely destroyed by human activities (Baylis 1958).

PLACOSTYLUS 14]

Previous Work on Placostylus bollonsi

Lesson & Martinet (1884: 227) referred to the existence of a species of Placostylus
recorded as Bulimus vibratus that was reputedly common at Three Kings Islands and that
differed from Placostylus on mainland Northland. The first known specimens collected
were obtained from above the landing in South East Bay, Great Island in April 1907, and
described as Placostylus bollonsi by Suter (1908). Haas (1935) subsequently placed the
species in an endemic monotypic subgenus Basileostylus. P. bollonsi was considered to
have become extinct by 1920, and searches of Great Island in 1928 and 1934 also failed to
locate any living snails (Fraser 1929; Powell 1935, 1938, 1948). However, during 1945-
46 three small extant colonies were found above South East and North West Bays and south
of Crater Head on Great Island, and larger populations were discovered on North East Island
in 1947, and on West Island in 1950 (Buddle 1948; Powell 1948, 1951). Snails from the
North West Bay and Crater Head colonies were found to have distinctive shell morphologies
and were described as P. bollonsi caperatus and P. bollonsi arbutus respectively by Powell
(1948), whereas the snails from South East Bay and North East and West Islands were
considered by Powell (1948, 1951) to be comparable with typical P. bollonsi. Powell
(1948, 1951) attributed the geographic distribution of subspecies of Placostylus on Three
Kings Islands to the effects of prehistoric human modification of vegetation cover, and to
inter-island transferrals by Maori. Climo (1973) discussed the distribution and taxonomic
relationships of P. bollonsi populations, described the reproductive system of the species,
and presented data showing that all three extant colonies on Great Island had undergone
increases in population size and range between 1946 and 1970.

METHODS
Population Structure

Size frequency data were obtained for P. bol/onsi populations on Great Island in March
1991. Four 25 m? quadrats were sampled in the areas occupied by each of the three colonies.
The distribution of habitat types containing live P. bollonsi snails was determined within
each colony area, and quadrats were randomly sited within those. Within each quadrat a
thorough search was made for all living and dead P. bollonsi. Shell lengths of those found
were measured on site, and specimens were then returned to the quadrat area; live snails
were replaced within broadleaf litter, whereas dead shells were scattered randomly
throughout the quadrat.

A distinction was made between living and dead specimens, and between specimens
possessing a labial varix and those with an unthickened outer lip. Callus deposition around
the apertural margin of P. bollonsi shells does not begin until an individual has almost
attained maximum shell length, and thus the presence of a varix is a useful criterion for
distinguishing ‘adult’ from ‘juvenile’ shell morphological types.

Qualitative observations on P. bollonsi colonies on North East and West Islands were
made during brief visits in 1982-83.

142 BROOK & LAURENSON

Population Size and Range

The geographic extent of each of the P. bollonsi colonies on Great Island was
determined in November 1983 and March 1991, by searching for live snails and shells of
recently dead snails, and recording their distribution on a 1:10 000 scale topographical map
(i.e. enlargement of Department of Survey and Land Information NZMS 270 Sheet L01).
Brief field notes were made on the distribution of gross vegetation types within and adjacent
to each colony, and abundances of live snails in habitats were qualitatively assessed.
Estimates of population sizes of the three colonies in 1983 and 1991 are made on the basis
of geographic areas occupied by snails, known and inferred densities of live snails as
determined from quadrats and qualitative observations, and assessment of the proportion
of habitat containing snails within each colony.

Morphometric Analysis

Randomly selected samples of 20 empty adult Placostylus shells retaining periostracum
and apertural colour, and with no obvious growth abnormalities, were made at sites within
each ofthe colonies on North East, West and Great Islands during 1982-83. Shells collected
from the same colonies in 1946-51, numbering 8-12 individuals per site, were examined
in collections held at the Auckland Institute and Museum, and the National Museum of New
Zealand, Wellington. The holotype of P. bollonsi Suter collected in 1907, and 9 shells
collected in 1946 from an extinct colony near Hapuka Point on Great Island were also
included in biometric analyses. Locality data for the samples are given in Appendix 1.

Ten linear measurements of shell dimensions of each of 160 shells were made to the
nearest 0.1 mm using vernier callipers (Fig. 3). These characters were selected to determine
differences in shell height, apertural height, spire angle and whorl inflation within and
between populations. All measurements were made by one person (FJB).

Principal components analysis was carried out on a correlation matrix of the total
multivariate data set using the PRINCOMP procedure in the SAS statistical package (SAS/
STAT users guide, 6.03 ed.; SAS Institute, Cary, North Carolina 1988). The significance
of temporal variation in shell height and apertural height between 1946-51 and 1982-83
samples from each of the five extant P. bo/lonsi populations was assessed using two-tailed
t-tests.

RESULTS
Population Structure
This section presents information on size frequency, density and mortality within P.
bollonsi populations on Great Island as obtained from quadrat studies and general
observations, and gives brief descriptions of populations on West and North East Islands.

Great Island Colonies

Size frequency data for live P. bollonsi on Great Island, as determined from searches
of four 25 m* quadrats within each colony in March 1991, are shown in Figs. 4-6. Adult

PLACOSTYLUS 143

Fig. 3. Dimensions of P. bollonsi shells used for morphometric analysis.

snails predominate within areas surveyed in the P. b. bollonsi and P. b. arbutus colonies
(74% and 67% of populations respectively), but were found to occur in subequal
proportions to juveniles within the P. 5. caperatus colony (i.e. 57% adults). Only one of
the individual 25 m* quadrats contained more juvenile than adult snails, but qualitative
observations made within all three colonies indicate that juveniles predominate locally in
peripheral areas.

Size frequency distribution data for P. bollonsi indicate that all populations on Great
Island are strongly normally skewed towards adult size classes. Data for P. b. bollonsi and
P. b. arbutus have a single large unimodal peak within the adult size range, whereas those
for P. b. caperatus show the same large unimodal peak of adult abundance, but also show
two smaller peaks of juvenile abundance at shell heights of 35-45 mm and 65-75 mm
respectively (Fig. 5). These juvenile peaks perhaps represent annual cohorts of P. b.
caperatus. If that interpretation is correct it indicates that juvenile shells increase in height
by c. 25-30 mm/year (i.e. from a hatchling height of about 11-16 mm), and attain adult size
at about 3 years old.

144 Brook & LAURENSON

P.b.bollonsi P.b.caperatus

no. of individuals

—
RS «ft

cS
my a wl wo ~ Lae 2 wn oo ™~

shell height (mm) shell height (mm)

no. of individuals

80
08
0
0

aS rear iY as
ae EG ota OS OE

shell height (mm)

Figs. 4-6. Size frequency distributions of shells of live P. bollonsi snails on Great
Island in March 1991, based on searches of four 25 m* quadrats within each of the
populations. Shaded and unshaded bars denote numbers of individuals with unthickened
and thickened outer lips respectively. 4. P.b. bollonsi. 5, P.b, caperatus. 6. P.b. arbutus.

Densities of live Placostylus snails recorded in quadrats on Great Island are listed in
Table 1. Snails are present at similar densities in the P. b. bollonsi and P. b. arbutus
colonies, but the P. b. caperatus colony differs in having considerably higher juvenile
densities, and hence higher overall snail densities than the other two colonies. Live snails
were typically sparsely and apparently randomly distributed within quadrats that had a
gently to moderately sloping substrate and an extensive cover of deep broadleaf litter.
However, on steeper slopes and in areas where the leaf litter cover was thinner or less
extensive, live snails were generally only present within pockets of broadleaf litter that were
stabilised within scree, or by groundcover plants such as Asplenium oblongifolium or Carex
testacea, and behind logs, and the roots and trunks of trees. In such areas the patchy
distribution of snails produces marked small scale variation in snail density. Localised
densities of up to 2-3 snails/m? were observed within small isolated leaf litter pockets in
some quadrats during the 1991 survey, but densities of up to 4-6 snails/m? were recorded
locally in comparable microhabitats within the P. b. caperatus and P. b. arbutus colonies
in 1982.

Searches of quadrats on Great Island in March 1991 failed to locate any live
Placostylus eggs, but rare empty eggs (i.e. lacking a cuticle and with a corroded surface)
and a few eggshell fragments were found. In contrast, single eggs were relatively common
throughout broadleaf litter within areas occupied by colonies on West and Great Islands in
December 1982, and on North East and Great Islands in November 1983, and Climo (1973)
noted that numerous eggs were present at each of the three Great Island colonies in
November 1970.

PLACOSTYLUS 145

Table |. Mean and range of densities of live P. bollonsi on Great Island as determined from
four 25 m? quadrats within each colony in March 1991.

P. b. bollonsi P. b. caperatus P. b. arbutus

Adult + mean + S.E. 0.18 + 0.03 0.34 + 0.14 0.15 + 0.05
juveniles/m? range 0.12 - 0.28 0.08 - 0.72 0.08 - 0.28
Juveniles/m? mean + S.E, 0.05 + 0.03 0.15 + 0.07 0.05 + 0.02
range 0 - 0.12 0 - 0.32 0 - 0.08
Adults/m? mean + S.E. 0.13 + 0.04 0.19 + 0.07 0.1 + 0.03
range 0.04 - 0.24 0.08 - 0.4 0.04 - 0.2

Size frequency data for spatially and temporally averaged populations of empty P.
bollonsi shells, as determined from searches of four 25 m? quadrats within each of the
colonies on Great Island, are given in Figs. 7-9. The three shell collections each show
similar bimodal distributions, with peaks in adult size classes and in the 15-20 mm juvenile
size class. Proportions of juvenile to adult shells are also similar within the three
populations comprising 56%, 58% and 53% of the P. b. bollonsi, P. b. caperatus and P. b.
arbutus populations respectively. Hatchling shells (10-20 mm size class) are almost
certainly under-represented in the collections as a consequence of their being weakly
calcified and very fragile compared with larger shells, and thus actual juvenile mortality
levels within the colonies on Great Island are probably considerably higher than the
assessed percentages.

7 P.b.bollonsi 8 P.b.caperatus

no. of individuals
no. of individuals

9 P.b.arbutus

no. of individuals

shell height (mm)

Figs. 7-9. Size frequency distributions of empty P. bollonsi shells based on searches of
four 25 m? quadrats within each of the P. bollonsi populations on Great Island in March
1991. Shaded and unshaded bars denote numbers of individuals with unthickened and
thickened outer lips respectively. 7. P.b. bollonsi. 8. P.b. caperatus. 9. P.b. arbutus.

146 BROOK & LAURENSON

No examples of P. bollonsi shells damaged by predators were seen in quadrats, or
elsewhere on Great Island.

West Island

In December 1982, live P. bollonsi were found at c. 100 m elevation on the ridge crest
at the southeastern end of West Island, and above c. 80 m elevation on the steep SW-facing
slope above the Elingamite wreck site. Elsewhere only scattered dead shells were seen.
There were at least 50 live individuals present at the southeastern end of the island, in a
terraced area under a 3 m high canopy of pukanui, karaka (Corynocarpus laevigatus) and
Three Kings rangiora (Brachyglottis arborescens). Live snails were found only under
Carex sedges in areas with deep, moist broadleaf litter, but empty shells were present both
under Carex and throughout adjacent areas of broadleaf litter. Some shells of freshly dead
adult snails were present in thin, dry litter at the margin of the grove. Living and dead P.
bollonsi juveniles were very rare at this locality, but solitary eggs were common in
broadleaf litter. No quantitative abundance data were obtained, but clumped densities of
up to 3-5 live snails per m? were observed locally. No information is available on the
abundance and density of snails on the steep southwestern side of West Island. In that area
rare live snails were present in ponded accumulations of broadleaf litter beneath a low
mixed broadleaf canopy. The overall population size of P. bollonsi on West Island is not
known, but it probably comprised at least 100 individuals in 1982.

North East Island

During a brief visit in November 1983, live P. bollonsi were found over an area of
about 1.5 ha across much of the forested top and upper slopes of North East Island. Snails
were scattered throughout litter and under Carex and groundferns, beneath a canopy
dominated by pukanui with some pohutukawa. Qualitative observations suggest that live
adult and juvenile snails were present in similar proportions and densities to P. b. bollonsi
and P. b. arbutus in broadleaf forest on Great Island (see above), indicating that the
population on North East Island probably comprised at least 200 individuals in 1983.

Colony Expansion on Great Island, 1945 to 1991

Comparison of 1991 data with those in Powell (1948) and Climo (1973) indicates that
all three extant P. bollonsi colonies on Great Island have increased substantially in size and
range since 1945 (Figs. 10-13). Temporal changes in size and extent of the P. b. bollonsi,
P. b. caperatus and P. b. arbutus colonies are described separately below.

P. bollonsi bollonsi Colony

Powell (1948: 283) states that in 1945 this colony covered an area of c. 20 m* ona
bouldery slope of about 45°, on the lower side ofa long rocky cliff face in moderately dense
scrub (see Fig. 10 for location). At that time a total of 11 live adult snails (and no live
juveniles) were recorded under the leaf spread area of 7 trees of wharangi (Melicope
ternata) and | of mahoe (Melicytus ramiflorus). By 1970 the colony comprised a population
of at least 30 live individuals (including juveniles) concentrated near a group of broadleaf
trees on the seaward side of the rock escarpment, but also extending sparsely in either

PLACOSTYLUS 147

North West Bay

South East Bay

0 metres 200
ee

= area occupied by populations in 1945-46
area colonised between 1945-46 and 1970
area colonised between 1970 and 1983

area colonised between 1983 and 1991

Location of presumed extinct Hapuka point population

Contour interval 50 metres

Figs. 10-11. Geographic ranges of P. bollonsi populations on Great Island in 1945-46,
1970, 1983 and 1991. The locations of these areas are shown in Fig. 1. 10. P.b. bollonsi
(upper right) and P.b. caperatus (lower left). 11. P.b. arbutus.

148 BROOK & LAURENSON

_8P.6.bollonsi

Pe I
=)
ti
—
)
D
=
o
_
—
=)
a,
2
=
a
r)
a

_— —A4P.b.caperatus

+ ——
1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995

year

Fig. 12. Estimated range increases of P. bollonsi populations on Great Island from
1945-46 to 1991.

P.b.arbutus

N wo w
o ) oi
o (=) o
t

®P.b.bollonsi

AP. b.caperatus

estimated minimum population size

1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995

year

Fig. 13. Estimated minimum sizes of P. bollonsi populations on Great Island between

1945-46 and 1991. Data for 1945-46 and 1970 are from Powell (1948) and Climo

(1973) respectively. Methods for estimating 1983 and 1991 population sizes are
outlined in the text.

PLACOSTYLUS 149

direction along the base of the escarpment for a total distance of about 60 m (Climo 1973:
585). Live snails apparently occurred over an area of about 0.12 ha at that time. In 1983
the colony was found to occupy an area of about 0.47 ha, with live snails occurring in mixed
broadleaf-kanuka forest for as much as 60 m downslope from the base of the escarpment,
and westwards to the ridge at the western end of the escarpment. The colony had also
extended to the top of the escarpment into kanuka dominated forest in the east (Fig. 10).
By 1991 P. b. bollonsi had sparsely colonised an extensive area of kanuka-dominated and
broadleaf forest in the valley to the north and east of the escarpment, and rare live snails
were also present in pukanui forest in a steep gully above the head of North West Bay. The
area occupied by the colony in March 1991 is estimated to be about 1.7 ha in extent (Figs.
10 and 12). Conservative qualitative assessments indicate that live snails occurred in about
20% of the total colony area occupied by 1983, but that only about 10% of the area colonised
between 1983 and 1991 is occupied owing to the patchiness of suitable microhabitats with
stable broadleaf litter or other ground cover. Thus live snails are inferred to have been
present over total areas ofc. 0.1 hain 1983 andc. 0.25 hain 1991. Ifthe assumption is made
that live snails were present within areas of suitable habitat at an average density of 0.1
snails/m? (i.e. conservative extrapolation from quadrat data - above), then it follows that
the estimated minimum size of the P. b. bollonsi population would have increased from c.
90 snails in 1983 to 210 snails in 1991 (Fig. 13). The maximum linear extension of the
colony periphery for the period 1970-1991 (excluding areas where snails may have rolled
or been washed downslope), was approximately 120 m, giving an average rate of increase
of about 5-6 m/year.

P. bollonsi caperatus Colony

In 1946 this colony comprised at least 40 living snails in an area of dense ngaio
(Myoporum laetum) scrub (Powell 1948: 283). Climo (1973: 586) described the colony as
being present in a valley on the seaward side of a low rocky escarpment and observed about
60 living snails. He noted that the snails were concentrated on the outer edge of the valley
near its higher end in shrubby broadleaf vegetation, but also extended sparsely into kanuka
forest adjacent to the escarpment, and down the valley slope. From the description given
in Climo (1973) it is estimated that the colony occupied an area of about 0.15 ha in 1970,
but by 1983 it had expanded to cover an area of about 0.3 ha (Figs. 10 and 12). Living snails
were still concentrated under broadleaf trees and shrubs on the seaward side of the valley
in 1983, but also extended 60 m SSW up a boulder escarpment and into mixed broadleaf-
kanuka forest and scrub at the head of the valley, and rare snails were present locally under
Carex sedges in kanuka-broadleaf scrub at the top of the rock escarpment bordering the
southeastern side of the valley. The colony subsequently expanded south from the head of
the valley into an area of kanuka- dominated forest with common broadleaf shrubs, and by
1991 covered a total area of c. 0.5 ha (Figs. 10 and 12). Conservative qualitative estimates
and extrapolations from quantitative quadrat studies (above), suggest that in 1983 and 1991
live snails were present over about 0.06 ha or 20% of the total colony area occupied within
the valley at an average density of about 0.2 snails/m*. In 1991 living snails are also
estimated to have been present over about 0.02 ha or 10% of the colony area to the south
of the valley at an average density of about 0.1 snails/m*. These estimates, if correct,
indicate that the minimum size of the P. b. caperatus population would have increased from
c. 110 snails in 1983 to 130 snails in 1991 (Fig. 13). Maximum linear extension of the colony
periphery for the period 1970-1991 was approximately 110 m, giving an average increase
of about 5 m/year.

150 BROOK & LAURENSON

P. bollonsi arbutus Colony

In 1945 this colony was restricted to an area of boulders at the foot of a W-facing scree
slope, beneath a smal] remnant grove of broadleaf trees (Powell 1948: 284). Twenty-five
living adult snails (no live juveniles) were observed in broadleaf litter ponded between
boulders, but Powell (1948) considered that more than twice that number of snails could
have been present. By 1970 the colony included at least 100 living specimens, and had
extended up the scree slope to a small cliff marking the upper limit of the valley, south into
kanuka forest on the edge of the broadleaf grove at the centre of the colony, and NW down
the slope from the rock scree into an area with scattered Cordyline kaspar trees (Climo
1973:585). Atthat time living snails apparently occurred over an area of about 0.3 ha (Figs.
11 and 12). The colony increased dramatically in size after 1970, expanding into kanuka-
broadleaf forest and shrubland on ridge crests to the north and east, and extending further
downslope to the west within mixed kanuka- broadleaf-Cordyline forest (Fig. 11). The P.
b. arbutus colony is estimated to have had a range of about 1.65 ha in 1983 and 2.7 ha in
1991 (Fig. 12). On the basis of qualitative observations, and extrapolation from quadrat
data (above), it is estimated that in 1983 and 1991 snails were present in about 20% of the
colony area centred on the remnant broadleaf grove (i.e. the 0.3 ha occupied by 1970), at
an average density of c. 0.2 snails/m*. Coverage and densities of snails elsewhere within
the colony in 1983 and 1991 were lower as a consequence of the patchiness of broadleaf
vegetation and areas of stable broadleaf litter. Live snails were conservatively estimated
to have been present in about 10% of the area colonised between 1970 and 1991, at an
average density of c. 0.1 snails/m?. On the basis of those estimates the minimum size of
the P. b. arbutus population would have increased from c. 250 individuals in 1983 to 360
individuals in 1991 (Fig. 13). Maximum linear extension of the colony periphery for the
period 1970- 1991 was approximately 240 m, giving an average rate of increase of about
11 m/year.

Morphometric Analysis

The results of principal components analysis (PCA) on a multivariate data set
incorporating the P. bollonsi Suter holotype, and samples of shells collected from P.
bollonsi colonies on North East, Great and West Islands in 1945-51 and 1982-83 are given
in Figs. 14-17. Cumulative eigenvalues of the correlation matrices for Prin 1 + Prin 2
describe 81% of variability within the data set. Significant features of this PCA plot with
respect to variation within the 1982-83 samples (i.e. Figs. 14-17) are:

(1) The 3 samples from extant Great Island populations form discrete, non-overlapping
albeit adjacent fields.

(2) The field defined by the North East Island sample is partly overlapping with those of
the P. b. bollonsi and P. 6. arbutus samples (cf. Powell 1948).

(3) The field defined by the West Island sample is partly overlapping with that of the P.
b. caperatus sample, but does not overlap with the P. b. bollonsi field (cf. Powell 1951).
(4) The P. bollonsi Suter holotype plots within fields defined by both the P. b. bollonsi
and North East Island samples.

(5) The field defined by 9 individuals from an extinct P. bollonsi colony at Hapuka Point
on Great Island is partly overlapping with fields defined by samples from each of the 3
extant populations on Great Island.

PLACOSTYLUS 15]

-3.5

* P.bollonsiSuter holotype; o Great Is. P.b.bollonsi;o Great Is. P.b caperatus; 4 Great Is. P.b.arbutus
vy West Is. P.bo/lonsi;© North East Is. P.bollonsi; x Hapuka Point P.bollonsi
Open symbols denote 1982-83 samples; closed symbols denote 1945-51 samples.

Figs. 14-17. Two dimensional configuration produced by principal components
analysis of a multivariate morphometric data set incorporating shell measurements
from the P. bollonsi Suter holotype, and from samples collected from P. bollonsi
populations in 1945-51 and 1982-83. 14. Samples collected from the five extant
populations in 1982-83. 15. Samples collected from the 3 extant Great Island
populations in 1982 and from an extinct population at Hapuka Point in 1946. 16. Pairs
of samples collected from extant Great Island populations in 1945-46 and 1982. 17.
Pairs of samples collected from populations on West and North East Islands in 1947-
51 and 1982-83.

152 BROOK & LAURENSON

A comparison of 1945-51 and 1982-83 PCA groupings for samples from each of the
5 extant populations (Figs. 16, 17) indicates that ranges of morphological variability shown
by the 1982-83 West Island, North East Island and P. b. caperatus samples are partly
overlapping with those of the respective 1945-51 samples. Data points for 7 out of 8 P. b.
bollonsi individuals from the 1945 sample are tightly clustered and enclosed within the
1982 P. b. bollonsi field, but the 8th data point is markedly disjunct from other P. b. bollonsi.
The 1945 P. b. arbutus field is completely enclosed within the 1982 field in terms of Prin
1 and Prin 2 values (Fig. 16), but shows only partial overlap with respect to ranges of Prin
3 values,

Summarised data and results of t-tests presented in Tables 2 and 3 indicate that there
are no significant differences in variance at < = 0.05 (two-tailed test) between either shell
height or apertural height values within samples collected in 1945-48 and 1982-83 from
each of the P. b. bollonsi, P. b. caperatus, P. b. arbutus and North East Island populations.
Differences in variance of shell height between 1950-51 and 1982 samples from the West
Island population are similarly not significant, whereas differences in variance of apertural

height between the two West Island samples are statistically significant at t, ,. ,, =+2.042

and t = 2.564. )
DISCUSSION

Life History

Placostylus bollonsi snails are hermaphroditic (Climo 1973). Casual observations of
temporal variability in abundance of eggs suggest the possibility that egg laying by P.
bollonsi is seasonal, occurring in spring and/or early summer, with most eggs hatching by
late summer. Timing of breeding, and duration of the period between laying and hatching
are not known. The large size of P. bollonsi eggs (12.5-18.0 mm long) and the fact that they
are apparently laid singly, indicate that egg production and laying behaviour of P. bollonsi
differ from those of Placostylus (Maoristylus) species. Snails in the latter group lay one
dozen to several dozen small eggs (5.7-6.6 mm long) in depressions or ‘nests’ covered by
10-15 mm of loose soil and leaf litter (Choat and Schiel 1980, Penniket 1981).

There is no direct information on growth rates of P. bollonsi but size frequency data
for P. b. caperatus show 3 evenly spaced size cohorts (4 including hatchlings), suggesting
the possibility that snails increase in length by about 25-30 mm/year and attain ‘adult’ size
at about 3 years old. In contrast, Maoristylus species are thought to reach adult size in 4-
6 years, albeit with growth rates showing positive correlation with the moisture content of
leaf litter (Penniket 1981). The rates of turnover of adult snails in P. bol/lonsi populations
are not known. However, given that numbers of snails in Great Island populations are
increasing rapidly despite the fact that those populations apparently contain only very low
proportions of juvenile size classes, it follows that the average age of adult snails 1s likely
to be at least as much as 8-10 years and could be significantly more.

Size frequency data for samples of empty P. bollonsi shells suggest that at least 60%
of snails fail to attain adult size and indicate that relative mortality rates of hatchlings are
significantly higher than those of larger juvenile size classes. Comparison of the size

PLACOSTYLUS 153

7090 =") 07 1'S#S°76 6 001-r'r8 861 (puejs] Iseq YON)
eega= 6 8 °7F9' £6 Z66-T16 —- SP-LHGT OIL] Nsuas ssuojJOq “q ‘d
o60'g =). 07 b'€FT'78 8°L8-8'bL 7861 (puejs] 1S9\)
wOwF= 4 ZI 0'P+6'P8 S06-SLL —_— IS-OS6T IR] Nsuas sswo]jJ0g “q ‘d
(0U1}x9) (‘] 1dIH Iq eyndepy)
6 C'S¥0'96 L'7OI-L'v8 9P61 OW] NSuas ssuO7]0q “q ‘d
6ST =") 0z l'bF1'SOl O'ST1-L'66 7861 (purjs] 18915)
grow Fa} 01 l'€Fr'Lol 6 E1I-P' E01 Sr6l SNINGAD “q ‘d
Lg =") 0z 5 €F1'€6 Z'ZOI-1'L8 7861 (pueys] 12215)
cTF= 4 a 0'PF6'16 0'L6-0'r8 9r61 snpsadbo “q ‘d
60r0= "3 0Z 6 €F1°T 8°86-£°S8 7861 (pueys] 301)
9907 F =" 4 8 1'C#S 16 PS6-L'88 —- (ZI S61-)S 61 1suo]jog °G ‘d
suvowi ‘dod jo soussayyip = “y aydures sod (uur)
"WOPdd1J JO SooISOp U SJeNplAIpul “G'S F1YSIOy (WW) S}YsIOY UONdI][O9

pue ‘sg’ =r 8 onfeA} = JO JOqUINNY « [Joys uUeopY =—s«JOYS Joosuey jo iva

‘suoryeyndod jue}xa ay} Jo yors wos sojdwes jounlsip A[jesoduss} Jo
SUROUI U99MI9q C()'() = 2 IL $}S9}-} POTIe}-OM] JO S}[NSoI pure ‘EQ-7RGH] PUR [S-SP6| Ul SpuL]s] SSury oIY L,
WO} Pojoa]]OO Ssajduies isuojjog ‘_q UIYM JWYSIOY [[9YS JO UONLIASpP psJepuLjs pue uvoW ‘asueY “7 9IqeL

154 BROOK & LAURENSON

9S10="1 02 b'1¥0°6€ O'1b-0°9€ €861 (pureys] seq YON)
OT F= 6 OTFI6E SZr9ISE Sb-LHGI IE] NSUDS ssUO]JOq “q ‘d
y9S'C = 4 02 CIFLEL LSe-6 1€ C861 (puejs] som)
COTF= I ZI GIFISE E8ESTE — IS-OS6T Ie] NsUAS ssuO]]0q “q ‘d
(Joulyx9) ~=—s ("| yeOIH ‘gq eyndey)
6 SCFLLE DIVO CE 9P6l OFe] NSUIS IsuUO71]/0g “G ‘d
$061 = “ 07 CCT IP CLY-S'LE C861 (purys] oI)
sp 7F= 01 OIF87 = 7'SP-6'IP Sol sningav “q ‘d
Se a ‘ 07 c FEV9e C BES EE C861 (puejs] yeo1p)
SpOT Fa" 4 a 6 0FS'SE 9°9€-S'°E€ 961 snypsadvo “q ‘df
1390 =99 0Z 9 1F8°8E E1p-7SE 7861 (puejs] 219)
9907 F = 3 8 9 OFT '6E LOr-rse — (eI s61-)sv61 1suoz]Og “4 ‘d
suvow ‘dod jo so0uasayy1p = “4 ayduies sod (wu) (wu) s}Yys19Y4
"WOpse1j JO SdoISOp U sjenplAipul “q's F#IYySsIOy = [eanyode UOT}DAT[OO
pure ‘cQ’9 = 38 oN[RA } joJoquinn yeinjyiode uvopy = jo odury JO 9X

‘suonejndod jueyx9 94} Jo yoeo wo sojdues younlsip AT[e1oduw19} Jo suvouw

UdIMI9Q S()'() = 2 J $}S9}-] PoyIe}-OM} JO S}[NSoI puke ‘¢Q-7R6] PUL [S-SP6| Ul SpuLls] ssury soIY], WO
payoayjoo soydures isuozjog ‘qd UIy{IM JYSIOY [eInjIode Jo UONeIASp plepue}s pure uesW ‘osuRY “¢ B[GRL

PLACOSTYLUS 155

frequency distributions of dead P. bollonsi with those of P. (M.) hongii populations from
areas without introduced mammalian predators (see data for Poor Knights Islands in Choat
& Schiel 1980, Penniket 1981), indicates that ‘natural’ levels of hatchling mortality are
markedly higher in the latter species. There appear to be fundamental differences in
recruitment strategies between the two New Zealand subgenera of Placostylus, with
Maoristylus producing numerous small eggs and suffering proportionately higher hatchling
mortality, whereas Basileostylus produces fewer and larger eggs, and has a proportionately
greater survival rate of resultant hatchlings.

There is no evidence that predation is a significant cause of mortality in P. bollonsi,
in spite of the presence of a number of potential predators of small juvenile snails at Three
Kings Islands (i.e. birds, reptiles, arthropods). Most juvenile mortality probably occurs as
a consequence of dessication of snails during dry periods, whereas adults probably die
mainly of dessication or old age.

Ecology

P. bollonsi snails occur in forest and shrubland habitats, and are ground-dwelling
herbivores. There is no record of the species occurring arboreally. From qualitative
observations it appears likely that the principal factors determining the distribution of snails
within colonies are the presence of broadleaf food plants, and of sheltered microhabitats
under broadleaf litter or groundcover plants (e.g. ferns, Carex spp., Muehlenbeckia
complexa). The preferred food plants of P. bollonsi have not been determined, but fallen
leaves of various broadleaf shrub and tree species, including pukanui, karaka, wharangi,
mahoe, Three Kings rangiora, ngaio, hangehange (Geniostoma rupestre var. ligustrifolium),
parapara (Pisonia brunoniana) and Coprosma macrocarpa are probably eaten. Live adult
and juvenile snails are typically sparsely and apparently randomly distributed in areas with
extensive, stable broadleaf litter, but are clustered in pockets of broadleaf litter or under
groundcover vegetation on steep slopes and in areas where broadleaf food plants are
sparsely or patchily distributed. P. bollonsi occurs at localised densities of up to 2-6 snails/
m? within such microhabitat patches, whereas mean densities determined from 25 m*
quadrats are much lower, ranging from 0.15-0.35 snails/m’, In contrast, P. (M.) hongii at
Poor Knights Islands is reported as occurring at overall mean densities of 2.11 snails/m* and
8.12 snails/m? on Tawhiti Rahi and Aorangi Islands respectively (Penniket 1981).

P. bollonsi snails under natural conditions are apparently active only at night,
remaining sheltered in cryptic microhabitats during daylight. Observations by Climo
(1973) suggest that dispersal of snails is random, and that individuals can cover distances
of up to several metres per night. Intensity of activity is apparently related to weather
conditions, being greater during or. following periods of rain. Juvenile snails locally
predominate in peripheral areas of colonies on Great Island, suggesting that they are
probably at least as active as adults, and play an important role in extending the boundaries
of colonies through emigration. Expansion of geographic ranges of the Great Island
colonies, and to a lesser extent parallel increases in their population sizes (Figs. 10-13), are
evidently in part limited by the nature of the vegetation cover in surrounding habitats.
Extensive broadleaf regeneration since 1946 in and adjacent to areas occupied by the P. 5.
bollonsi and P. b. arbutus colonies (e.g. Cameron et al. 1987), has presumably made it
possible for those populations to undergo rapid range expansion, whereas the spread of P.
b. caperatus has probably been limited by a persistence of kanuka-dominated vegetation

156 BROOK & LAURENSON

in the surrounding area, Changes in areas occupied by colonies on Great Island since 1970
indicate that P. bollonsi snails are capable of crawling up rock escarpments and steep, bare,
friable soil slopes, and of moving into small isolated patches of broadleaf vegetation within
areas of kanuka forest or shrubland. However, their successful dispersal across areas of
inhospitable habitat is probably limited to wet periods when snails would be able to escape
dessication. Given the rapid range expansion of colonies, and the ability of P. bollonsi
snails to disperse across areas lacking broadleaf vegetation cover or stable litter, Climo’s
(1973) assertion that the three extant subspecies on Great Island are unlikely to intermingle,
is probably incorrect.

Variation in Shell Morphology

Results of the morphometric analysis presented above are subject to the following
known and inferred sources of error. Because of the relatively small sample sizes used in
this study (8-20 individuals per population), it is likely that the range of morphological
variation determined for each is less than that actually present. However, the relative
location and size of each of the sample clusters on the PCA plot do give an indication of
relative degrees of morphological variability within and between populations. The samples
are time-averaged in that they comprise solely dead collected (1982-83), or a mixture of
live and dead collected (1945-51) individuals. An attempt was made to minimise any
temporal bias by including only shells of relatively freshly dead snails (shells with
periostracum and apertural colouring) in the analysis. The statistically significant
difference between apc.“ 1ral heights of the 1950-51 and 1982 West Island samples are
probably caused by non-iandom sampling resulting from the preferential collection of large
shells in 1950-51 (see Powell 1951: 132), The fact that the three extant Great Island colonies
contained only very low numbers of snails/empty shells in 1945-46 (Powell 1948) indicates
that collections made from them at that time are probably reasonably representative of
actual contemporary morphological variability. However, the markedly disjunct grouping
of individuals in the 1945 P. b, bollonsi sample as determined by PCA (Fig. 16) suggests
the possibility that the morphologically distinct individual within that sample (an unlabelled
shell in AK 25995, see Appendix 1) may in fact have been collected from North East Island
(cf. Figs 14, 17).

Considering the above shortcomings within the data set, morphometric analysis
supports Powell’s (1948) contentions that the three extant P. bollonsi populations on Great
Island have differing shell morphologies, and that shells from the extant P. b. bollonsi
population above South East Bay are morphologically comparable with the P. bollonsi
Suter holotype. Furthermore, it is apparent that characteristic shell morphologies have
persisted within each of the populations from 1945-46 to 1982-83, although there is some
indication that the overall range of morphological variation may have increased slightly in
the P. b. bollonsi and P. b. arbutus populations during that time.

The analysis indicates that the range of morphological variation within the extinct
Hapuka Point population is not directly comparable with that in the P. b. arbutus population
as claimed by Powell (1948), but rather overlaps with those of all three extant populations
on Great Island. Given the small size of the Hapuka Point sample (9 individuals) and wide
spacing of its data points on the PCA plot, it is likely that the actual morphological variation
within the original population was even greater.

PLACOSTYLUS 157

Powell (1948, 1951) considered that shells from the West Island and North East Island
populations were morphologically most similar to those from the P. b. bollonsi population
on Great Island, but that is not supported by the present analysis. The West Island samples
show a range of morphologic variation that is most similar to that of P. b. caperatus, albeit
with shells from the former population typically being slightly smaller and having lower
spire ratios (cf. Powell 1951; Tables 2, 3). The overall range of morphological variation
within samples from the North East Island population overlaps with that of samples from
both the P. b. bollonsi and P. b. arbutus populations on Great Island, although as pointed
out by Powell (1948), the North East Island samples are most similar to P. b. bollonsi in
terms of average shell height and spire ratio.

Differences in shell morphologies between geographically isolated populations of a
land snail species can result from direct environmental influences on snail growth and shell
shape, and/or from genetic differentiation by adaptive or non-adaptive selection, or by
random processes (e.g. Jones et al. 1977; Goodfriend 1986). The differences in shell
morphology between P. bollonsi populations are considered to be largely genetically
controlled rather than phenotypes of environmental origin for two reasons. First, each of
the three extant populations on Great Island has apparently retained distinctive and separate
morphological characteristics from 1945-46 to 1982. If local environmental conditions
were important in determining differences in shell morphology between the three small and
locally restricted 1945-46 populations, it is unlikely that those shell characteristics would
have continued to dominate in descendant populations, given the significant environmental
changes that occurred on Great Island following the removal of goats in 1946. Probably
as a consequence of their very limited geographic ranges, each of the 1945-46 populations
were subjected to relatively homogeneous environmental settings in terms of substrate and
vegetation type, litter cover and slope aspect (see Powell 1948). However, post-1946
expansion of ranges and population sizes coupled with marked albeit patchy habitat
changes in vegetation and litter cover, meant that each of the colonies subsequently came
to occupy a much wider range of environmental settings. That leads to the second reason
for assuming that the morphological variation is largely genetic, that although the P. b.
bollonsiand P. b. arbutus populations on Great Island each presently occupy a diverse albeit
physically comparable range of habitat types (vegetation cover, substrate types, slope
aspect, climate), the observed morphological differences between the two populations
exceed intrapopulation variation. If environmental factors have an important inductive
effect on P. bollonsi shell shape, the ranges of morphological variation of those two
populations should be more comparable than they are.

Although the morphological characteristics of the extant P. bollonsi populations are
here considered to be predominantly genetic, it is possible that some of that differentiation
originally occurred as a result of environmental selection. In a review of the nature and
causes of variation in size and form of land snail shells, Goodfriend (1986) noted that with
many species, individuals living in moist conditions tend to be larger and have greater
relative aperture areas than individuals in drier habitats. There is no comparable
relationship evident within existing P. bollonsi populations between shell form and the
various habitat types in which the species occurs, as noted above. However, such a
correlation may have existed on Great Island in 1945-46, with individuals from the two
populations that were restricted to boulder scree under remnant broadleaf groves (i.e. P. b.
bollonsi and P. b. arbutus) typically having larger relative aperture areas and broader spires

158 BROOK & LAURENSON

than the P. b. caperatus individuals which were restricted to a (presumably) drier habitat
under ngaio shrubs (see Powell 1948).

Although the observed differences in shell morphologies between the various P.
bollonsi populations are probably largely genetic, a recent electrophoretic study by Triggs
and Sherley (in press) based on samples of 5 snails from each of the North East Island and
three Great Island populations, found only minor genetic differentiation between populations.
That study showed that in terms of the 28 allozyme loci identified, the P. b. bollonsi, P. b.
caperatus and North East Island populations were virtually genetically identical, and that
the P. b. arbutus population differed from the other 3 only in terms of gene frequency. A
further conclusion was that P. bollonsi has a low level of genetic polymorphism and
heterozygosity.

Studies of some other landsnail species have also failed to find a consistent correlation
between molecular differentiation and observed heritable shell characters within different
populations (e.g. in Cepaea : Jones et al. 1977). In such cases, the nature of any molecular
differentiation between populations may need to be resolved at the DNA level.

Biogeography and Morphological Differentiation

The genus Placostylus is widely distributed on continental and volcanic islands of
Mesozoic and Cenozoic age in the southwest Pacific Ocean (Solem 1959), which, given
a presumed limited ability to disperse across ocean basins, suggests that it had an eastern
Gondwana origin. At present there are two subgenera of Placostylus in northern New
Zealand that differ in anatomy and shell morphology (Haas 1935; Solem 1959; Climo
1973). The subgenus Maoristylus includes two closely related species, P. (M.) hongii
Lesson in eastern Northland, and P. (M.) ambagiosus Suter in the Cape Reinga-North Cape
area (Powell 1979), and the monotypic Basileostylus incorporates the Three Kings Islands
endemic P. (B.) bollonsi.

There is no direct information on the Mesozoic-Cenozoic biogeography and phylogeny
of Placostylus in New Zealand, with the oldest known fossils of the genus being of late
Holocene age (Millener 1981). However, given what is known about the palaeogeography
of the northernmost New Zealand region, Placostylus ancestral to the present-day P.
bollonsi populations probably colonised the Three Kings Islands area in early Miocene
time, and were subsequently isolated on the insular land masses that have persisted in that
area from mid Miocene time to the present. During that period of about 15 Ma before the
arrival of Maori at Three Kings Islands, there was probably a complex history of genetic
differentiation within and between Placostylus populations in the area resulting from
environmental factors such as changes in climate, habitat types, and the distribution and
size of islands. It is logical to assume that an ancestral P. bollonsi population(s) would have
been widely distributed across a single large island that existed on Three Kings Bank during
the ultimate glacial period (i.e. prior to 11-14 Ka) and that subpopulations of that ancestral
population could subsequently have become isolated on the four larger islands that formed
within the Three Kings group during the period of post- glacial sea level rise. There were
possibly inherited genetic differences between some or all of these isolated populations
resulting from localised and/or clinal genetic differentiation within the ancestral
population(s), and differentiation probably occurred within each island population as a
consequence of major environmental changes (in climate and habitat types) caused by the
transition from glacial to interglacial conditions.

PLACOSTYLUS 159

Given the evidence for widespread anthropogenic habitat modification at Three Kings
Islands during several centuries before 1840, it is not definitely known if the 5 surviving
and | extinct populations of P. bollonsi all originated in situ, or whether the present-day
locations of one or more resulted from translocations by Maori. Previous studies based on
shell shape suggested that the North East Island population originated in prehistoric
transfers of P. b. bollonsi snails from Great Island (Powell 1948, Climo 1973), and Powell
(1951) proposed a similar origin for the West Island population. The morphometric
analysis presented above shows clearly that shell shapes in the West Island population differ
from those in any of the other P. bollonsi populations. Given the conclusion that shell shape
within P. bollonsi is predominantly genetic, the simplest hypothesis is that the West Island
‘race’ differentiated in situ. It is impossible to determine from the morphometric data
whether the North East Island population originated in situ, or resulted from transfers by
Maori. The latter is a possibility given that the ranges of shell shape within the two
populations are overlapping (not identical as previously claimed), but the simplest
hypothesis on the basis of shell shapes alone is that the North East Island population, and
those on Great Island, originated in situ.

During the period of Maori occupation of Three Kings Islands, widespread forest
clearance probably isolated small remnant populations of P. bollonsi in steep, rocky areas
on Great and North East Islands (Powell 1948), and possibly accounts for the absence of
P. bollonsi on South West Island. In the early part of the 19th century until about 1840,
P. bollonsi populations on Great Island were probably also adversely affected by domestic
goats and pigs, and a subsequent reintroduction of goats to that island in 1889 led to severe
habitat modification and consequent extinction or reduction in sizes of remnant P. bollonsi
populations (Baylis 1948, Powell 1948, Turbott 1948). After eradication of goats in 1946
broadleaf trees and shrubs on Great Island regenerated widely and the three remaining P.
bollonsi populations increased substantially in size and range. Similar increases in snail
numbers probably also occurred on Great Island between 1840 and the 1890s in the absence
of human inhabitants and before the build- up of goat numbers, and the snail population on
North East Island had probably expanded to cover much of the formerly cultivated top of
that island by the late 19th or early 20th century (Buddle 1948, Baylis 1958).

Possibly much of the observed morphological differentiation between P. bollonsi
populations on Great Island is owing to “founder effects” resulting from population
bottlenecks during the periods of Maori occupation and goat infestation. The fact that the
now extinct Hapuka Point population had a range of morphological variation that
substantially overlaps with those of the three extant populations on Great Island, lends
support to this hypothesis. However, at least some of the differentiation between Great
Island populations may have been inherited from a pre-Maori population(s), and/or have
occurred in response to environmental factors during the last few hundred years. The range
and size of the North East Island population were probably also significantly lower during
the period of Maori occupation than at present, suggesting the possibility that it too could
have been subjected to random selection processes resulting from population bottlenecks.
West Island is apparently the only large island in the Three Kings group on which broadleaf
forest and shrub communities escaped extensive modification (Baylis 1958), and it is
unlikely that size, range or genetic diversity of that P. bollonsi “race” were ever signifi-
cantly affected by anthropogenic activities.

160 BROOK & LAURENSON

Conservation Status

Historic records show that before the removal of goats from Great Island in 1946 the
three P. bollonsi populations on that island were in danger of becoming extinct following
progressive degradation of habitat caused by the browsing and trampling of goats (Powell
1948). Subsequent recovery of those snail populations is directly attributable to habitat
improvement resulting from regeneration of broadleaf vegetation in the absence of goats.
It is thus obvious that one of the most serious potential threats to the continued survival of
P. bollonsi populations at Three Kings Islands would be renewed destruction or adverse
disturbance of areas of broadleaf vegetation by fire, or from the introduction of browsing
mammals or fungal pathogens (see comment by P.J. Brook in Wright & Cameron 1990).
The other main potential threat is the introduction of mammalian predators such as rats or
pigs, given the marked adverse effect those animals have had on Placostylus populations
elsewhere (see Powell 1938, 1979; Iredale 1944; Penniket 1981). In the continued absence
of such habitat modification the West Island and North East Island P. bollonsi populations
will probably persist with similar numbers of snails as are present now, whereas the three
populations on Great Island should continue to expand in size and range, and will
presumably eventually amalgamate assuming broadleaf regeneration on that island
continues. None of the P. bollonsi populations require any management intervention at
present, but it would clearly be advantageous to acquire more detailed information on
growth rates, population turnover, and ecological requirements of the species, should such
intervention become necessary in the future.

Acknowledgements. We are grateful to the following people: Judy Roberts for typing the manuscript
and Terry Conaghan for drafting the figures, Conway Subritzky (m.v. Pandora), Eric Gosse (m.v.
Pegasus IT) and Ken Durey (m.v. Harrold Harvey) for transport to Three Kings Islands in 1982, 1983
and 1991 respectively. Peter Brook and John McCallum for assisting with fieldwork in 1982. Bruce
Marshall, Walter Cernohorsky and Bruce Hayward for access to collections held at the National
Museum and Auckland Institute and Museum. Jenny Riley for helping record some of the biometric
data. Chris Green for providing information on the Three Kings Islands terrestrial arthropod fauna.
Malcolm Francis for giving advice on biometric analysis and running the SAS principal components
analysis programme. Peter de Lange, Richard Parrish, Ray Pierce, Greg Sherley and Kath Walker
for suggesting improvements to the manuscript.

REFERENCES
Bavtis, G.T.S.
1948 Vegetation of Great Island, Three Kings Group. Rec. Auckland Inst. Mus. 3:
239-252.

195la £lingamita (Myrsinaceae) a new monotypic genus from West Island, Three Kings.
Rec. Auckland Inst. Mus. 4: 99-102.

1951b Incipient forest regeneration on Great Island, Three Kings group. Rec, Auckland
Inst. Mus. 4: 103-109.

1958 A botanical survey of the small islands of the Three Kings group. Rec. Auckland Inst.
Mus. 5: 1-12.

1986 Widened seed dispersal at the Three Kings Islands attributed to naturalised birds. Jn:
A.E. Wright & R.E. Beever (Eds.) The offshore islands of northern New Zealand.
Department of Lands & Survey Information Series 16: 41- 45.

PLACOSTYLUS 161

Brook, F.J.
1989 Sheet N1 and N2 North Cape and Three Kings (1st Ed.). Geological map of New
Zealand 1:63 360. DSIR, Wellington, New Zealand.

Brook, F.J. and G.P. THRASHER
1991 Cretaceous and Cenozoic geology of northernmost New Zealand. N.Z. Geol. Surv.
Rec. 41: 1-42.

Bubb e, G.A.
1948 The outlying islands of the Three Kings group. Rec. Auckland Inst. Mus. 3: 195-204.

Cameron, E.K., G.T.S. Bay is and A.E. WRIGHT
1987 Vegetation quadrats 1982-83 and broad regeneration patterns on Great Island, Three
Kings Islands, northern New Zealand. Rec. Auckland Inst. Mus. 24: 163-185.

CarTeErR, R.M., C.A. Lanpis, J.J. Wittiams and L. CARTER
1983 Post-glacial evolution of the Otago continental shelf. Pacific Science Association
15th congress. Programme, abstracts and congress information 1: 35.

CHEESEMAN, T.F.
1888 Notes on the Three Kings Islands. Trans. & Proc. N.Z. Inst. 20: 141-150.
1891 Further notes on the Three Kings Islands. Trans. & Proc. N.Z. Inst. 23: 408-424.

Cuoat, J.H. and D.R. ScHiEL
1980 Population structure of Placostylus hongii (Gastropoda: Paryphantidae) on the Poor
Knights Islands. N.Z. J. Zool. 7: 199-205.

Cito, F.M.
1973 The systematics, biology and zoogeography of the land snail fauna of Great Island,
Three Kings group, New Zealand. J. R. Soc. N.Z. 3: 565-628.

CRANWELL, L.M.
1962 Endemism and isolation in the Three Kings Islands, New Zealand - with notes on
pollen and spore types of the endemics. Rec. Auckland Inst. Mus. 5: 215-232.

Doak, W.
1969 = The Elingamite and its Treasure. Hodder and Stoughton, Auckland. 192p.
FALLA, R.A.
1948 Anew Anthornis from Three Kings Islands. Rec. Auckland Inst. Mus. 3: 337-338.
Forster, R.R.
1954 The New Zealand harvestmen (suborder Laniatores). Canterbury Mus. Bull. 2: 200-
207.
FraAseR, W.M.

1929 Notes on a visit to the Three Kings Islands. N.Z. J. Sci. & Tech. 11: 148-156.

Gps, J.G.
1986 A New Zealand regional Holocene eustatic sea-level curve and its application to
determination of vertical tectonic movements. Roy. Soc. N.Z. Bull. 24: 377-395.

162 BROOK & LAURENSON

Gites, E.T.
1958 Dermaptera from the Three Kings Islands, New Zealand, with the description
of a new species of Brachylabis Dohrn (Labiduridae). Rec. Auckland Inst. Mus.

5: 43-48.
Given, B.B.
1954 Melolonthinae (Coleoptera) from the Three Kings Islands. Rec. Auckland Inst. Mus.
4: 267-270.

GOoDFRIEND, G.A.
1986 Variation in land-snail shell form and size and its causes: a review. Syst. Zool. 35:

204-223.
Haas, F.
1935 Beschreibung neuer Untergattungen und Arten von Mollusken. Zoolog. Anzeiger
109: 188-195.
Haywarp, B.W.

1987 Prehistoric archaeological sites on the Three Kings Islands, northern New Zealand.
Rec. Auckland Inst. Mus. 24: 147-161.

Haywarpb, B.W. and P.R. Moore
1987 Geology of Three Kings Islands, northern New Zealand. Rec. Auckland Inst. Mus.
24: 215-232.

HeEereEs, J.E.
1898 Abel Jonszoon Tasman’s Journal. Amsterdam, Muller and Co. 417p.

Ho.psworth, M.
1951 Effect of goats on Great Island, Three Kings: the permanent quadrats resurveyed.
Rec. Auckland Inst. Mus. 4: 113-121.

Hotpswortn, M. and G.T.S. Bay is
1967 Vegetation of Great Island, Three Kings group, in 1963. Rec. Auckland Inst. Mus.
6: 175-184.
Hotioway, B.A.
1963 Wing development and evolution of New Zealand Lucanidae (Insecta: Coleoptera).
Trans. R. Soc. N.Z. 3: 100-116.

IREDALE, T.
1944 The land Mollusca of Lord Howe Island. Australian Zool. 10: 229-334.

Jones, J.S., B.H. Leite and P. RAWLINGS
1977 Polymorphism in Cepaea: a problem with too many solutions? Ann. Rev. Ecol. Syst.
8: 109-143.

LABILLARDIERE, J.J.
1799 Relation du voyage a la recherche de la Perouse. Tome second. 113p.

Leacu, H.M.
1984 1000 years of Gardening in New Zealand. Wellington, Reed.

PLACOSTYLUS 163

Lee, K.E.
1959 The earthworm fauna of New Zealand. New Zealand Department of Scientific and
Industrial Research Bulletin 130: 1-387.

Lesson, P.A. and L. MARTINET
1884 Les Polynésiens. Vol. 4.

McCann, Y.H.C.
1955 The lizards of New Zealand. Gekkonidae and Scincidae. Dominion Mus. Bull. 17: 1-
127.

McNas, R.
1908 Historical Records of New Zealand. Volume 1. Wellington. 779p.

Mare es, B.J.
1956 Spiders from the Three Kings Islands. Rec. Auckland Inst. Mus. 4: 329- 342.

MILLENER, P.R.
1981 The Quaternary avifauna of the North Island, New Zealand. Unpublished Ph.D.
thesis, University of Auckland.

OLIver, W.R.B.
1948 The flora of the Three Kings Islands. Rec. Auckland Inst. Mus. 3: 211-238.

PENNIKET, A.S.W.
1981 Population studies of landsnails of the genus Placostylus in the north of New
Zealand. Unpublished M.Sc. thesis, University of Auckland.

PoweELL, A.W.B.
1935 Land mollusca of the Three Kings Islands, New Zealand. Proc. Mal. Soc. London 21:
243-248.

1938 The Paryphantidae of New Zealand, No.4, and the genus Placostylus in New Zealand.
Rec. Auckland Inst. Mus. 2: 133-150.

1948 Land Mollusca of the Three Kings Islands. Rec. Auckland Inst. Mus. 3: 273-290.

1951 Land Mollusca from four islands of the Three Kings group, with descriptions of three
new species. Rec. Auckland Inst. Mus. 4: 127-133.

1979 New Zealand Mollusca. Collins. 500Op.

Puckey, W.
1836 Missionary register for 1836, p.473. London.

Roth, H.L.
1891 Crozet’s Voyage to Tasmania, New Zealand, etc. London, Truslove & Shirley. 148p.

SALMON, J.T.
1948a Collembola from the Three Kings Islands with a description of Proisotomina, new
genus. Rec. Auckland Inst. Mus. 3: 291-300.
1948b New genera, species and records of Orthoptera from the Three Kings Islands, New
Zealand. Rec. Auckland Inst. Mus. 3: 301-308.
1948c New species and records of Lepidoptera from the Three Kings Islands, New Zealand.
Rec. Auckland Inst. Mus. 3: 309-312.

164 BROOK & LAURENSON

SoLeM, A.
1959 Systematics and zoogeography of the land and freshwater mollusca of the New
Hebrides. Fieldiana : Zool. 43: 1-359.

STEVENS, G., M. McGLone and B. McCuLtocu
1988 Prehistoric New Zealand. Auckland, Heinemann Reed. 128p.

Suter, H.
1908 Anew Placostylus from New Zealand. Trans. & Proc. N.Z. Inst. 40: 340- 343.

Tricas, S.J. and G.H. SHERLEY
In press. Allozyme genetic diversity within Placostylus landsnails and implications for
conservation. N.Z. J. Zool.

Tursott, E.G.
1948 Effects of goats on Great Island, Three Kings, with descriptions of vegetation
quadrats. Rec. Auckland Inst. Mus. 3: 253-272.

Woopwarp, T.E.
1950 A new species of Cermatulus Dallas from the Three Kings Islands, New Zealand
(Heteroptera:Pentatumidae). Rec. Auckland Inst. Mus. 4: 24-30.
1954 New records and descriptions of Hemiptera-Heteroptera from the Three Kings
Islands. Rec. Auckland Inst. Mus. 4: 215-234.

Waricut, A.E. and E.K. CAMERON
1990 Vegetation management on northern offshore islands. Jn: Towns, D.R.; Daugherty,
C.H.; Atkinson, I.A.E. (Eds.). Ecological restoration of New Zealand Islands.
Conservation Sciences Publication 2: 221-239.

F.J. BROOK, Department of Conservation, P.O. Box 842, Whangarei. C.-M. LAURENSON,
7 Kokich Crescent, Whangarei.

APPENDIX 1.

PLACOSTYLUS 165

Locality information for Placostylus bollonsi shells used for biometric

and morphological analysis. AK = Auckland Museum collections, M = National Museum
collections. (n) = number of specimens used for analysis. LO1/ = grid references from
NZMS 260 series maps (Department of Survey and Land Information).

a i SE TR

Placostylus bollonsi bollonsi, Great Island

Holotype, National Museum (1)

AK 25061 (3)
AK 26288 (1)
AK 25995 (3)
AK 76365 (1)
AK 75135 (20)

Placostylus bollonsi caperatus, Great Island

AK71095, Holotype (1)

(10)
(20)

AK 76366
AK 75137

Placostylus bollonsi arbutus, Great Island

AK 71094, Holotype (1)

AK 76367 (9)
AK 75133 (10)
AK 75134 (10)

South East Bay landing slope. April 1907.

No. | location, % mile NE of provision depot (ie.
= No. 2 location of Powell 1948:283), November-
December 1945.

500' in valley of SE landing (no date given,
presumably 1945-1951).

No. | location, “4 mile NW (ie. should be NE) of
provision depot (see above), November- December
1945.

No. 1 location, NE of provision depot (see above),
November-December 1945.

North side of South East Bay at c. 100 m elevation.
L01/322832. December 1982.

Approximately 375' elevation, NW landing slope,
April 1946.

As above.

Cliffs above North West Bay at c. 130 m elevation.
L01/319830. December 1982.

No. 2 location, SW coast (ie. = No. 5 location of
Powell 1948:284); no date given, presumably
November-December 1945.

As above.

W of trig station at c. 240 m elevation. L01/308827.
December 1982.

Ridge NW of trig station at c. 260 m elevation. LO1/
30808275. December 1982.

Placostylus bollonsi sensu lato, Great Island (extinct)

AK 25062 (9)

Placostylus bollonsi sensu lato, West Island

AK 25064 (3)
AK 25065 (3)
AK 25066 (3)

Hapuka Point, Great Island (see Powell 1948:284),
April 1946.

West Island, 1951.
West Island, 1951.
West Island, December-January 1950.

166 BROOK & LAURENSON

AK 26482 (3)
AK 75136 (20)

West Island (no date giver, presutnably 1950 or
1951),

Tetrace at c. 100 m elevation at SE etd of island.
L01/222793. December 1982.

Placostylus bollonsi sensu lato, North East Island

AK 76364 (5)
M 5853 (1)
M 5859 (3)
AK 75132 (20)

North East Island (no date given, presutttably 1947
or 1948).

North East Island, ex Aucklatid Museum (no date
given, presumably 1947 or 1948).

North East Island, January 1948.

Top of island at c. 90 tn elevation. L01/34258535.
Novetnbetr 1983.

THE NEW ZEALAND SPECIES OF BEROTHIDAE
(INSECTA: NEUROPTERA)

K.A.J. WISE

AUCKLAND INSTITUTE AND MUSEUM

Abstract. The one described New Zealand species of Berathidge, Profahiella zelandica
Tillyard, 1923, was previously known from anly 2 female specimens. Additional
records are nated. A new description, based on male and female specimens, is given.
It is concluded that P. zelandica stands alone as 4 separate endemic species and is the
only New Zealand Berothid.

The New Zealand Berothid lacewing Protabiella zelandica was described by Tillyard
(1923) from two specimens. One, which he collected on 4 February 192] in the Nelson area.
was designated as the holotype. He designated the other specimen as a paratype. Jt was
collected at Waitati an 13 December 1916, probably by C.E. Clarke, an amateur
entomologist of Dunedin, who callected another lJacewing at Waitati on 1] December 1916
(Micromus tasmaniae Walk., C.E. Clarke collection, Auckland Museum),

G.V. Hudson, of Wellington, callected the species several times in a valley near
Wellington between 192] and 1925. After his collection on 6 December 1923, he wrote a
nate the next day describing the living insect (Hudson 1924), Later (Hudsan 1950), he gave
further similar information apparently based on observations in the same valley on 13
January 1938, byt no specimens of this date have heen found in collections.

Navas (1929), in his revision of the Family Berothidae, included short descriptions of
both Pratohiella and P. zelandica based on the descriptions and figure of Tillyard (1923).
Wise (1963, 199]) gave knawn distribution limits, Wise (1977) listed the species and
MacLeod & Adams (1967) and Oswald & Penny (1991) listed the genus.

Tillyard (1923) had designated his two specimens as 4 weet male and 4 paratype
female, but Aspéck & Aspick (1985) and Aspdck (1986) found that both specimens are
females and re-descrihed and figured this sex. Male specimens have now heen discavered
sa the species is here re-described and figured (Fig. |), and collecting data from all known
specimens are given.

Specimens are in the collections of Auckland Institute and Museym (AMNAZ); British
Museum (Natural History), London, England (BMNH); Canterbury Museum, Christchurch
(CMNZ); National Museum, Wellington (NMNZ); and Landcare Research (previously
Plant Protection Entomology, DSIR), Auckland (NZAC).

Rec, Auckland Inst. Mus. 29: 167-177 1992

‘(PAPIPPIT “f 19d) “e810H Aqkem ‘o[eUlay “EZHI ‘PAeAT[LL vIpunjaz vyjaIqgolo4g “| “314

Ay"()

[i bye iy Zp, A NZ / aA avail }
TN ih Wal (( MANA Vis rs ei, ay AA Wy Me) ) HM AME iy

168 WISE

BEROTHIDAE 169

FAMILY BEROTHIDAE
Subfamily BEROTHINAE
Genus Protobiella Tillyard, 1923

Protobiella Tillyard, 1923, Trans. Proc. N.Z. Inst. 54: 218 (NZ).

Protobiella ; Navas, 1929, Mems Acad. Cienc. Exact. Fis.-Quim. Nat. Zaragoza 2: 90 (NZ).
Protobiella ; MacLeod & Adams, 1967, Psyche 74(3): 258 (NZ).

Protobiella : Aspéck & Aspéck, 1985, Z. ArbGem. Osterr. Ent. 36(3, 4): 75, 80 (NZ).
Protobiella : Aspéck, 1986, Recent Research Neuropterology, 89-92, 100 (NZ).

Protobiella : Oswald & Penny, 1991, Occas. Pap. California Acad. Sci. No. 147: 50 (NZ).
Protobiella ; Wise, 1991, Rec. Auckland Inst. Mus. 28: 212, 223 (NZ).

Protobiella zelandica Tillyard, 1923 (Figs. 1-10)

Protobiella zelandica Tillyard, 1923, Trans. Proc. N.Z. Inst. 54: 219 (NZ).

Protobiella zealandica : Hudson, 1924, Ent. Mon. Mag. 60: 40 (NZ), [in error for zelandica].

Protobiella zelandica : Tillyard, 1926, Insects Australia New Zealand, 316 (NZ).

Protobiella zelandica : Navas, 1929, Mems Acad. Cienc. Exact. Fis.-Quim, Nat. Zaragoza 2: 91
(NZ).

Protobiella zelandica : Hudson, 1950, Fragments New Zealand Entomology, 116 (NZ).

Protobiella zelandica : Wise, 1963, Pacific Ins. 5(1): 54 (NZ).

Protobiella zelandica : Wise, 1977, Bull. Auckland Inst. Mus, 11: 132 (NZ).

Protobiella zelandica : Aspéck & Aspéck, 1985, Z. ArbGem. Osterr. Ent, 36(3, 4): 75 (NZ).

Protobiella zelandica : Aspock, 1986, Recent Research Neuropterology, 98 (NZ).

Protobiella zelandica : Wise, 1991, Rec. Auckland Inst. Mus, 28: 212, 214 (NZ).

Small brown species with yellow-brown markings on head and thorax, showing
laterally as interrupted stripe. Antennae of medium length, setose, dark brown but pale
apically. Head and body with black setae. Wings speckled with brown, darker brown spots
showing as a transverse subapical stripe and two short inner posterior stripes on anterior
wings. Underside of body and legs pale brown, legs darker apically with a distinct pale mark
basally on tarsus.

Antennae with elongate basal segment (scape) ca. 3 times long as wide, 36-45 flagellar
segments, dark brown basally with apical ca. 12 segments very pale. Flagellar segments -
oo" : 36-39, 99 : 41-45. Head shining brown with yellow markings, face-mouthparts yellow
except clypeus-lower edge of frons with transverse dark brown band. Dorsally head brown
anteriorly extending between antennae (top of frons) with a dark brown wart on each side,
dorsum dark brown, yellow laterally; posteriorly with 2 transverse non-setose warts.

Thorax shining brown with yellow markings. Pronotum slightly elongate with flange
on each side, dark brown with thin pale yellow median stripe, the yellow lateral stripes
showing dorsally as stripes or spots. Mesothorax dark brown dorsally with elongate yellow
spot on each side anteriorly. Metathorax brown with pale median area. Legs setose (? more
so in females); pale brown, darker apically on tibia and tarsus; basal tarsal segment long
with basal half paler showing as a distinct pale mark.

170 wise

Wings with membrane mainly clear, veins setose, edges thickly setose; pterostigma
not delineated, area thickly setose. Forewing broad; apex rounded, slightly tapering in
males well-rounded in females; membrane with series of brown marks along costa, brown
spats elsewhere at vein forks and cross-veins with in some specimens extra spots in
transverse stripes; main veins with subequal light and dark dashes; setae unicolorous, dark,
except small patches of pale (ginger) setae on costa between adjacent dark membrane spots,
small patches of pale (straw to golden-straw) setae variously around apex and posterior
margin. Hindwing narrow, tapering apically; membrane clear without spots; setae unicolorous
brown except some small pale (straw to golden-straw) patches around margin.

Wing venation (Figs. 2, 3). Forewing with Costa (C) bowed near base, costal area
broadest at ca. ’4 of wing, humeral vein simple but variable, many forked veinlets,
pterostigma not differentiated. One basal cross-vein Subcosta-Radius (Sc-R), Sc fused with
R apically (Sc+R), with forked veinlets at margin. Remainder of veins also with forked
veinlets at margin. Radial sector, Medius anterior (Rs,Ma) partially fused, with Ma basad
from Rs+Ma, several branches of Rs apicad. Three cross-veins R-Rs. Branches of Rs
variable in number from 4 or more in males to 7 in females, also with occasional variations
(see Fig.3, forewing Rs5); with cross-veins in gradate series to Ma variable. Base of Ma
present as apparent cross-vein (b) at Rs+Ma -Medius posterior (Mp), apical cross-vein Ma-
Mp. Mp forked, with cross-vein closing fork; 3 cross-veins M, Mp - Cubitus (Cu). Cu arises
separately with 2 branches Cubitus anterior, Cubitus posterior (Cua, Cup), with cross-vein
closing fork; 1 cross-vein Cup - Anal (A). Three A (Al, A2, A3) arise separately, with 1
crass-vein Al-A2.

Hindwing with costal area narrow, many simple veinlets, pterostigma not differentiated.
No cross-veins Sc-R, Sc fused with R apically, with forked veinlets to margin. Remainder
of veins also with forked veinlets to margin. Rs, Ma partially fused, with Ma basad from
RRst+Ma, several branches of Rs apicad. Two cross-veins R-Rs, Branches of Rs variable in
number, 4 or more in males, to 6 in females, no gradate cross-veins. Base of Ma present
as apparent cross-vein (b) at Rs+Ma-Mp, | apical cross-vein Ma-Mp in some (? females).
Mp forked with cross-vein closing fork; | apical cross-vein Mp-Cua. Cua arises separately,
| cross-vein Cug-A; 2 or3 A present. Rs branches, M, Cua noticeably long, slightly curved,
syb-parallel to each other and to posterior margin.

Abdomen brown dorsally with darker median stripe, various darker patches on 6 basal
segments, pale straw-colour ventrally extending up posterior edges of basal segments,
terminal segments straw-colour.

Genitalia (Figs. 4-10). 0’. Tergite VIII (T VIII) little differentiated, narrow. T IX fused
with ectoproct (e), elements visible, rounded on both sides. Torulus absent. Sternite VIII
(S VIII) little differentiated, narrow. S IX separate with short setae; coxopodite (c) elongate,
thick plate, straight-sided, trilobed apically, hinges upwards against ectoproct; paramere-
mediuncus complex (pm) ventrally in coxopodite in median line, almost parallel-sided,
slightly dilated apically, apex truncate, with 3 or 4 long, pointed setae. Gonarcus (g) with
2 elongate, upcurved processes (gp) medially. Hypandrium internum (hi) triangular.

9. Tergite VII (T VII) little differentiated; T VIII elongate laterally. T IX fused with
ectoproct (e). Ventral genital area anteriorly with Sternite VII (S VII) little differentiated;

BEROTHIDAE 171

possesses , Sco#A ;
Dos eo DY LD LP,
nyt eee
by. fe ' yt L ye ice Rs
%

PEE SG
Cf

yy qe

WEED

Fig. 2, 3. Protobiella zelandica Tillyard, 1923. Wing venation. 2.0%. 3.9. (For
abbreviations see description). (Del. J. Liddiard).

172 WISE

Fig. 4-10. Protobiella zelandica Tillyard, 1923. Genitalia. 4-7.0°. 4. Lateral. 5. Dorsal.

6. Ventral. 7. Dorso-caudal. 8-10.9. 8. Lateral. 9. Ventral. 10. Spermatheca and bursa

copulatrix to exterior. (For abbreviations see description). (Del. 4, 6-10
P. Quinn, 5 K.A.J.W.).

BEROTHIDAE 173

S VII small, pointed, setose, with rounded transverse pudiculum (pd) posteriorly. T [X+e
elongate laterally; a pair of large finger-like processes (pseudohypocaudae - Aspock &
Aspock 1985, ph) projecting ventrally; gonapophyses laterales (gl) elongate, without
hypocaudae. Spermatheca (s) with 2 thick coils, one through other, a thin tube to large bursa
copulatrix (bc).

Measurements. Anterior wing: o’o’ 6.75-9.0 mm, $9 7.0-9.0 mm. Posterior wing: oo”
6.0-7.5 mm, 29 6.0-7.75 mm.

Specimens examined. South Island: Holotype ?. Takaka, 6.11.1921,R.J. Tillyard. Protobiella
zelandica Till. Type o& R.J.T. (Tillyard 1923 - Terakohe, near Takaka, Golden Bay).
Paratype ¢. [Data label missing.] Protobiella zelandica Type ¢ R.J.T. (Tillyard 1923 -
Waitati, 13.XII.1916). Fletchers Ck, West Inangahua SF, BR, 25.1.1972, J.S. Dugdale, H.P.
McColl. o.

All 3 specimens are in NZAC and they are the only South Island specimens available.

North Island: Wayby Gorge, 30.XII.1926. C.E. Clarke collection. Protobiella zelandica

det. K.A.J.W. 1959.9. Ohakune, -.1.1927, T.R. Harris.c’.

53a, 53b (G.V. Hudson. Beaten out of forest growth, Gollans Valley, 24.XIJ.1921).
lo 1.

53c (G.V.H. Beaten out of forest growth, Gollans Valley, 25,1.1922). Gollans Valley,
Wellington, 25.1.1922, G.V. Hudson. Protobiella zelandica Till.o Det. by R.J.
Tillyard.@.

53d, 53e, 53g (G.V.H. Beaten out of silver tree ferns, Upper Gollans Valley, 6.XII.1923),
Gollans Valley, Wellington, 6.XIJ.1923, G.V. Hudson. B.M.1931-328. 2 oo" 1 9.

53f (G.V.H. Beaten out of silver tree ferns, Upper Gollans Valley, 6.XI1.1923).o”.

53h(G.V.H, Ditto [Beaten out of silver tree ferns, Upper Gollans Valley] 1.11.1924. Gollans
Valley, 1.11.1924, G.V. Hudson. B.M.1933-274.0".

53k (G.V.H. Beaten out of forest growth, Gollans Valley, 20.1.1925). Gollans Valley,
24.X11.1921 [in error], G.V. Hudson, B.M. 1932-218.9.

531 (G.V.H. Beaten out of forest growth, Gollans Valley, 20.1.1925). 1.

53m (G.V.H. Ditto [Beaten out of forest growth, Gollans Valley] 23.X1I.1925). Gollans
Valley, 6.XII.1923 [in error], G.V. Hudson. B.M. 1931-328.0".

53n, 53p, 53q, 53r (G.V.H. Beaten out of dead leaves of Cordyline indivisa, Erua,
31.X11.1931). 2 oo’ 19 1.

123 (G.V.H. Beaten from dead leaves of Cordyline indivisa, Erua, 31.XII1.1931). ide.
31.X11.1931, G.V. Hudson. B.M. 1932-218. 1 0” 2 99.

The G.V. Hudson labels are tiny number/letter (species/specimens) labels and the data
given above in brackets is from the G.V,. Hudson collection register in NMNZ; data from
other labels (where present) follows. Two printed BMNH labels have been given dates of
earlier collections in error. North Island specimens are in AMNZ, BMNH, CMNZ and
NMNZ.

174 wise

Type localities (Fig. 11). New Zealand: South Island. The holotype @ label data is “Takaka”
but it is recorded (Tillyard 1923) from ““Terakohe, near Takaka, Golden Bay”, which is
undoubtedly Tarakohe on the Golden Bay coast ca. 10 km from inland Takaka. The
paratype 9 data is recorded (Tillyard 1923) as Waitati which is ca. 15 km north of Dunedin.

Distribution (Fig. 11). One specimen has been collected north of Auckland and several in
the central North Island but most have been taken near Wellington. Two specimens are from
the north-western corner of the South Island and one from the south-eastern. The South
Island male is labelled from “Fletchers Ck” in West Inangahua State Forest. Fletcher Creek
is ca. 10 km south of Inangahua and is ca. 150 km southwards from Tarakohe.

Life history. The elliptical, white eggs, without sculpture, are supported at the end of a
long filament, singly or more often in a cluster on one filament (Hudson 1924). Hudson
(1924,1950) also observed that the brown insect rests with the antennae close together and
the tips appressed to the substrate which gives the appearance of a small dead leaf,
particularly when on a twig.

Gourlay (1930), Clark (1932) and Valentine (1967) have recorded Protobiella
zelandica as a predator but no specimens have been found to support these records and they
are possibly based on misidentification. All the references listed by Valentine (1967: 1147),
including Gourlay (1930), appear to arise from the work of A.F. Clark who recorded the
species (Clark 1932) as an infrequent predator of Adelgids in pine plantations. Recently
cleared abdomens of two adult specimens had what appeared to be pieces of plant material
in the gut; the larva is not yet known to the present author.

DISCUSSION

Protobiella zelandica appears to have a secondary sexual character in the number of
antennal flagellar segments. In specimens so far seen, males have less than 40 and females
more than 40.

Aspock & Aspock (1985) related Protobiella zelandica Tillyard, 1923 to
Austroberothella rieki Aspéck & Aspoéck, 1985 of Australia. The wings of both species are
figured in Aspock & Aspock (1985: 83) and both the colour pattern and wing venation are
noticeably different. Those authors, however, related the female genitalia of P. zelandica
to that of A. rieki, being of the same pattern with pseudohypocaudae and without
hypocaudae; the males of both species were not then known.

In general pattern, the male genitalia are similar to those of Trichoma gracilipenne
Tillyard 1916, as figured by Aspéck & Aspdéck (1985). The obvious torulus, seen in the
figures of male genitalia of three Australian species which they described, has not been seen
in males of P. zelandica but a median pair of elongate gonarcal processes occurs in this
species and not in the others. These differences, together with differences in the female
genitalia, may indicate a separation at a supra-generic level. Later, Aspock & Aspéck
(1988) described and figured the male genitalia of Austroberothella rieki which are quite
distinct and do not appear to be particularly close to those of Protobiella zelandica.

BEROTHIDAE 175

N
eWayby x

AUCKLAND

NORTH

ISLAND

Eruae
eOhakune
Golden Bay
Takaka arakohe 9
Ly
xe a
lelso ollans Valley
e Inangahua
SOUTH
ISLAND /,
\
p\ CHRISTCHURCH
4 B
jw
Q Zz FI Waitati
> DUNEDIN Pie cee ee ee

Fig. 11. Map of New Zealand with localities for Protobiella zelandica Tillyard,
1923. (Del. R. Gilbert).

176 WISE

One male of Protobiella zelandica from the South Island has been compared with
males from the North Island and there is no obvious difference to suggest that there is more
than one species. So far as is known this species stands alone as a New Zealand endemic
in the Family Berothidae.

Acknowledgements. Loans of specimens were made by Mr. Stephen Brooks, British Museum
(Natural History), London, England; Mr Ron Ordish, National Museum of New Zealand, Wellington;
Mr Anthony Savill, Canterbury Museum, Christchurch and Dr Trevor Crosby, New Zealand
Arthropod Collection, Landcare Research, Auckland. Information was provided by Dr John Ward,
Canterbury Museum and Dr Jean Legrand, Muséum National d'Histoire Naturelle, Paris, France.

Thanks are also given to Mr Jurgen Herbke, of Auckland, for translation from German; Miss
Joanna Liddiard, Middlesex Polytechnic, Barnet, England, and Mr Peter Quinn, Auckland, for
figures; and to Miss Rosemary Gilbert, Auckland Museum, for continuing assistance with
specimens, figures and maps.

REFERENCES

Aspock, U.
1986 The present state of knowledge of the Family Berothidae (Neuropteroidea: Planipennia).
pp. 87-101. In Recent Research in Neuropterology. Proc. 2nd. Int. Symp.
Neuropterology. J. Gepp, H. Aspéck, H. Hélzel (Eds.). Graz. 176p.

Aspock, U., and H. Asp6ck
1985 Die Berothiden Australiens (und Neuseelands) II: Die Genera Trichoma Tillyard,
Trichoberotha Handschin, Protobiella Tillyard und Austroberothella n.g.
(Neuropteroidea: Planipennia: Berothidae). Z. ArbGem. Osterr. Ent. 36(3, 4): 65-85.
1988 Die Berothiden Australiens V: Zur systematischen stellung von Austroberothella rieki
U.A. & H.A. (Mit einem kurzen Uberblick iiber die Erforschung der Berothidae
Australiens) / (Neuropteroidea: Planipennia). Stapfia 17: 135-146.

Crark, A.F.
1932 Insects infesting Pinus radiata in New Zealand. N.Z. J. Sci. Tech. 13(4): 235-243.

GovurLay, E. S.
1930 Preliminary host-list of the entomophagous insects in New Zealand. Bull. N.Z.D.S.L.R.
22: 1-13.

Hupson, G. V.
1924 Note on the habits of Protobiella zealandica Tillyard. Ent. Mon. Mag. 60: 40-41.
1950 Fragments of New Zealand Entomology. Wellington, Ferguson & Osborn. 188p.

MacLeonp, E. G., and P. A. ADAms
1967 A review of the taxonomy and morphology of the Berothidae, with the description of a
new subfamily from Chile (Neuroptera). Psyche 74(3): 237-265.

Navas, R.P.L.
1929 Monografia de la Familia de los Berdtidos (Insectos Neurépteros). Mems Acad. Cienc.
Exact. Fis.-Quim. Nat. Zaragoza 2: 1-107.

BEROTHIDAE 177

Oswa _p, J.D., and N.D. PENNY
1991 Genus-group names of the Neuroptera, Megaloptera and Raphidioptera of the world.
Occas. Pap. California Acad. Sci. 147: 1-94.

TILLYARD, R. J.
1923 Descriptions of new species and varieties of lacewings (Order Neuroptera Planipennia)
from New Zealand, belonging to the Families Berothidae and Hemerobiidae. Trans.
Proc. N.Z. Inst. 54: 217-225.
1926 The Insects of Australia and New Zealand. Sydney, Angus & Robertson. 560p.

VALENTINE, E. W.
1967 _ A list of the hosts of entomophagous insects of New Zealand. N.Z. J. Sci. 10(4): 1100-
1209.

Wisk, K. A. J.
1963 A list of the Neuroptera of New Zealand. Pacific Ins. 5(1): 53-58.
1977. Asynonymic checklist of the Hexapoda of the New Zealand sub-region. The smaller
Orders. Bull. Auckland Inst. Mus. 11: 1-176.
1991 Distribution and zoogeography of New Zealand Megaloptera and Neuroptera. Rec.
Auckland Inst. Mus. 28: 211-227.

K.A.J. WISE, c/o Auckland Institute and Museum, Private Bag 92018, Auckland.

.
ee Pa
~~

tah ne

See rae ae Behe
Lt tet aie re,

hig gal a aE

a iy a

Pelee: ay

oe ad
Fi

oll
<dhite mates teak %

ge ait Ait Bt
ifosearh: Desay aig

perenne >
ee
Peri, oun 4

= “4 Jeo =

: ae ae

RECORDS OF THE AUCKLAND INSTITUTE AND MUSEUM

Instructions for Authors

The Records contain the results of original research dealing with material in the collections of the Museum,
and research carried out by Museum staffmembers in their particular subjects. All papers are refereed. Submission
of a paper is taken to mean that the results reported have not been published or submitted for publication elsewhere.
Before submitting a manuscript, authors must consult a recent volume of the Records and ensure that their
manuscript follows the correct layout, conventions and format for sectional headings, tables, captions, references,
abbreviations, etc, Manuscripts not submitted in the required format may be returned without consideration. An
acceptable guide for style, other than recent volumes of the Records, is the latest edition of the New Zealand
Government Printing Office Style Book.

Manuscripts should be typed double-spaced on one side only of good quality white International A4 sheets,
with margins of at least 25 mm left and right. Draft-quality output from dot-matrix printers is hot acceptable. Two
copies of the text are required. Two photocopies of each photograph or other illustration should be supplied when
the paper is submitted. The original illustrations should be provided after the paper is accepted.

Authors should indicate if their manuscript is available in machine-readable form on standard word-
processor discs. Once the paper has been refereed, edited and accepted, the author will be asked to update the
computer file with all necessary changes and submit the corrected disc and one hard copy of the corrected
manuscript.

Abstracts are required for all papers and should give principal findings rather than describe the work done.
All measurements should be in metric (SI) units. Entries listed under “References” must be referred to in the text.
See recent volumes for the format and style of the reference list.

Figures and tables must be referred to in the text and their approximate positions should be marked. Figures
and groups of figures should be mounted, and tables composed, to page proportions with due regard to saving space.
They must be supplied ready for reduction (where necessary) to the printed page size or less (allowing for captions
- above tables, below figures), preferably upright on the page. Individual figures should be numbered consecutively
in arabic numerals (not la, b, c etc.) regardless of whether they are grouped or individual, or whether photographs
or line drawings. They should be numbered in order of reference.

Figure captions should be typed as a separate list. Tables, with captions above, should be submitted on
separate pages. If the author can generate high-quality tables or appendices on a laser printer, they may be treated
as figures and their captions listed separately.

Authors are responsible for the accuracy of their data, references and quotations, for their conclusions and
remarks, and for correcting proofs. Sole authors will receive 50 free offprints; joint authors 100 offprints between
them. Further offprints may be purchased.

a
Meee

Seal

mas

are MS

: “

A Tes,

4 et
Ar 2. ‘

&
Dv. >

—-
4 7
ee .

ees
gyi pe

; J Fe

. ae
ane tye as
a3 eat a bh

~s Rood

es
wm

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Texts

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "Records of the Auckland Institute and Museum"

See other formats